Targeting GRPR for sex hormone-dependent cancer after loss of E-cadherin

Raymond, Jérémy H.; Aktary, Zackie; Pouteaux, Marie; Petit, Valérie; Luciani, Flavie; Wehbe, Maria; Gizzi, Patrick; Bourban, Claire; Decaudin, Didier; Nemati, Fariba; Martianov, Igor; Davidson, Irwin; Tomasetto, Catherine-Laure; White, Richard M.; Mahuteau-Betzer, Florence; Vergier, Béatrice; Larue, Lionel; Delmas, Véronique

doi:10.1038/s41586-025-09111-x

Download PDF

Article
Open access
Published: 11 June 2025

Targeting GRPR for sex hormone-dependent cancer after loss of E-cadherin

Jérémy H. Raymond ORCID: orcid.org/0000-0002-8712-895X^1,2,
Zackie Aktary^1,2,
Marie Pouteaux^1,2,
Valérie Petit^1,2,
Flavie Luciani^1,2,
Maria Wehbe^1,2,
Patrick Gizzi³,
Claire Bourban³,
Didier Decaudin⁴,
Fariba Nemati⁴,
Igor Martianov⁵,
Irwin Davidson ORCID: orcid.org/0000-0001-5533-1171⁵,
Catherine-Laure Tomasetto⁵,
Richard M. White⁶,
Florence Mahuteau-Betzer^7,8,
Béatrice Vergier⁹,
Lionel Larue ORCID: orcid.org/0000-0002-2010-6107^1,2^na1 &
…
Véronique Delmas ORCID: orcid.org/0000-0001-7368-3664^1,2^na1

Nature volume 643, pages 801–809 (2025)Cite this article

11k Accesses
3 Citations
75 Altmetric
Metrics details

Subjects

A Publisher Correction to this article was published on 14 July 2025

This article has been updated

Abstract

Sex inequalities in cancer are well documented, but the current limited understanding is hindering advances in precision medicine and therapies¹. Consideration of ethnicity, age and sex is essential for the management of cancer patients because they underlie important differences in both incidence and response to treatment^2,3. Age-related hormone production, which is a consistent divergence between the sexes, is underestimated in cancers that are not recognized as being hormone dependent^4,5,6. Here, we show that premenopausal women have increased vulnerability to cancers, and we identify the cell–cell adhesion molecule E-cadherin as a crucial component in the oestrogen response in various cancers, including melanoma. In a mouse model of melanoma, we discovered an oestrogen-sensitizing pathway connecting E-cadherin, β-catenin, oestrogen receptor-α and GRPR that promotes melanoma aggressiveness in women. Inhibiting this pathway by targeting GRPR or oestrogen receptor-α reduces metastasis in mice, indicating its therapeutic potential. Our study introduces a concept linking hormone sensitivity and tumour phenotype in which hormones affect cell phenotype and aggressiveness. We have identified an integrated pro-tumour pathway in women and propose that targeting a G-protein-coupled receptor with drugs not commonly used for cancer treatment could be more effective in treating E-cadherin-dependent cancers in women. This study emphasizes the importance of sex-specific factors in cancer management and offers hope of improving outcomes in various cancers.

GRHL2 motif is associated with intratumor heterogeneity of cis-regulatory elements in luminal breast cancer

Article Open access 08 June 2022

Precision therapeutics and emerging strategies for HR-positive metastatic breast cancer

Article 23 August 2024

Deep learning from HE slides predicts the clinical benefit from adjuvant chemotherapy in hormone receptor-positive breast cancer patients

Article Open access 30 August 2021

Main

Cancer, which is one of the leading causes of premature mortality in humans, is known to have multiple risk factors. Some of these factors are intrinsic to the tumour or cell of origin, but many are extrinsic, originating from the tumour environment or individual behaviour. Sex and age have various effects on tissue and tumour exposure in cancer biology, but their role and, most importantly, their interactions remain poorly understood. Epidemiological data have revealed numerous sex differences in the incidence of various cancers in non-reproductive organs, although the role of sex hormones is sometimes suspected but little studied^1,6,7. Understanding these sex- and age-related dynamics is essential for improving prevention and treatment strategies.

E-cadherin as an oestrogen regulator

Despite men having an overall higher cancer risk than women, our sex- and age-stratified analysis of global epidemiological data reveals a higher risk of cancer incidence among premenopausal women (Fig. 1a and Extended Data Fig. 1a). Of 24 cancer types analysed, 13 showed a significant female bias during this period (P = 0.0248; Extended Data Fig. 1a–f and Supplementary Table 1), with malignant melanoma being the most prominent (Fig. 1b and Extended Data Fig. 1c). Melanoma occurs most frequently in pregnant women and those exposed to high oestrogen levels during puberty^8,9,10. This coincides with peak oestradiol levels in young women, whereas testosterone remains stable across ages, underscoring the potential role of oestrogen in initiating cancer (Fig. 1c and Extended Data Fig. 1g).

**Fig. 1: E-cadherin is a crucial hub in oestrogen-mediated cancer responses with sex-dependent effects on melanoma metastasis.**

Genes central to the main female-biased cancers (melanoma, gastric and thyroid), breast cancer and ESR1, which encodes oestrogen receptor-α (ERα), include CDH1 (which encodes E-cadherin), CCND1 (cyclin D1), BRAF and KRAS (Fig. 1d and Supplementary Table 2). Of these, only CDH1 mRNA levels were lower in young women than in men or older women (Fig. 1e), indicating that reduced CDH1 expression could increase oestrogen-driven cancer susceptibility, particularly in melanoma.

E-cadherin loss drives sex-biased metastasis

To investigate the role of E-cadherin in female-biased melanoma, we developed a melanoma mouse model with conditional Cdh1 deletion in melanocytes (Tyr::CreA/°; Cdh1^F/F). No melanoma developed within two years, indicating that Cdh1 loss alone is insufficient for tumour initiation. To induce melanoma, we combined Cdh1 loss with the NRAS(Q61K) mutation using the Tyr::NRAS^Q61K/°; Cdkn2a^+/− model¹¹. We chose NRAS over BRAF because there is more frequent CDH1 downregulation in NRAS-mutant melanomas and a lack of targeted therapies for NRAS-mutated cases (Extended Data Fig. 1h).

We monitored melanoma development in Tyr::NRAS^Q61K/°;Cdkn2a^+/−;°/°;Cdh1^F/F (Ecad) and Tyr::NRAS^Q61K/°;Cdkn2a^+/−;Tyr::CreA/°;Cdh1^F/F (∆Ecad) mice. Cdh1 loss did not affect melanoma onset, primary tumour count or penetrance (Fig. 1f). However, lung metastasis was significantly higher in female ∆Ecad mice: 63% (15 of 24) had metastases compared with 16% (3 of 19) of Ecad female mice (Fig. 1g–i). Both micro- and macro-metastases were more frequent in female ∆Ecad mice compared with other groups (Fig. 1j and Extended Data Fig. 1i,j). These findings indicate that Cdh1 loss enhances melanoma metastasis in a sex-dependent manner, particularly affecting female mice.

Ecad loss upregulates GRPR expression

RNA-seq analysis of eight primary tumours per genotype and sex revealed a marked upregulation of gastrin-releasing peptide receptor (Grpr) in ∆Ecad female melanomas compared with all other groups, including ∆Ecad male mice (Fig. 2a, Extended Data Fig. 2a and Supplementary Tables 3–6). H3K27ac chromatin immunoprecipitation followed by sequencing (ChIP–seq) confirmed there were active Grpr promoter regions exclusively in ∆Ecad female melanoma cell lines (Extended Data Fig. 2b), with expression maintained in female lung metastases (Extended Data Fig. 2c). This upregulation was independent of the Tyr::Cre line or the Cdh1 melanoma model (Extended Data Fig. 2d).

**Fig. 2: Loss of *Cdh1* Induces *Grpr* expression.**

The role of GRPR in promoting metastasis was supported by: first, its association with elevated invasive and neural crest cell-like gene signatures in human melanoma (Fig. 2b,c and Extended Data Fig. 2e–h); second, its reduced overall and progression-free survival in women with high GRPR levels, a trend not seen in non-sex-stratified data (Fig. 2d,e and Extended Data Fig. 2i,j); third, ECAD⁻/GRPR⁺ being the dominant phenotype in human lung melanoma metastases, regardless of genotype (Fig. 2f–i); and finally, the expression of GRP, which is GRPR’s natural agonist, in human and mouse lungs (Extended Data Fig. 2k–m).

GRPR mRNA is widely expressed across tumours, including melanomas¹². Notably, its expression is inversely correlated with CDH1 in both skin cutaneous melanoma (SKCM) and acral melanoma (ACRM), as well as in several carcinomas (breast, lung, stomach and thyroid), according to TCGA data and ref. ¹³ (Fig. 2j). This indicates that E-cadherin may commonly regulate GRPR across cancer types.

GRPR drives lung metastases and is targetable

The expression of GRP (the endogenous ligand of GRPR) in rodent and human lungs indicates that GRPR activation may drive metastasis-supporting mechanisms. We dissociated primary tumours and injected the cells into C57BL6/J mouse tail veins. Only ∆Ecad melanoma cells expressing Grpr colonized lungs within 30 days, unlike Ecad-expressing cells (Extended Data Fig. 3a). When we engrafted melanoma tumours, ∆Ecad female tumours showed significantly faster growth than Ecad male, Ecad female and ∆Ecad male tumours (Extended Data Fig. 3b), indicating the cell-autonomous aggressiveness of ΔEcad female melanomas. Cell lines from these tumours confirmed Grpr production uniquely in ΔEcad female lines (Supplementary Table 7).

To test the role of GRPR in metastasis, we attempted Grpr knockout in ∆Ecad melanoma cells and ectopic expression in Ecad cells. We could not knock down or knock out Grpr in female ∆Ecad cells (Extended Data Fig. 3c–i), implying that GRPR is essential in the absence of E-cadherin. However, we successfully introduced Grpr into male mouse 1181 and human 501mel cell lines. Grpr expression and activation in these engineered cell lines closely mirrored that of endogenous Grpr⁺ lines, such as 1057 and Dauv-1 (Extended Data Fig. 3j–m). Tail-vein injection of parental, control and Grpr-expressing cells demonstrated that GRPR expression alone is sufficient to promote lung metastasis (Fig. 3a–c). In both male and female NSG mice, 501mel cells ectopically expressing GRPR formed lung metastases, consistent with sex-independent expression driven by the CMV promoter.

**Fig. 3: *Grpr* expression and activation drive lung melanoma metastasis.**

To evaluate GRPR inhibition in vivo, we used luciferase-expressing 1057 ΔEcad female melanoma cells (1057Luc). After tail-vein injection, numerous Dct- and Grpr-positive lung metastases formed (Fig. 3d,e). Two GRPR antagonists, RC-3095 and PD-176252, had similar in vitro effects (Extended Data Fig. 3n,o), but RC-3095 was used owing to its higher metabolic stability (Extended Data Fig. 3p,q).

Mice injected with 1057Luc cells were randomized into vehicle or RC-3095 treatment groups (Extended Data Fig. 3r,s). Without treatment, thoracic luminescence appeared by day 24 and grew exponentially (Extended Data Fig. 3t). Targeting GRPR with RC-3095 significantly reduced lung colonization, as measured by luminescence and metastasis counts (Fig. 3f–i and Extended Data Fig. 3u). These results underscore the essential role of GRPR in lung metastasis formation, validated through both gain-of-function and pharmacological approaches.

GRPR fuels key metastasis pathway

To uncover the cellular mechanisms by which GRPR activation by GRP promotes metastasis, we first assessed its effect on growth. In line with the in vivo data, ∆Ecad female mouse melanoma cell lines gained colony-forming ability (Fig. 4a). Ectopic GRPR expression in mouse and human melanoma lines strongly induced (in 1181) or enhanced (in 1014 and 501mel) colony formation (Fig. 4b,c and Extended Data Fig. 4a). GRP stimulation promoted growth in GRPR-positive mouse (1057 and 1064) and human (MDA-MB-435S) cell lines, whereas co-treatment with GRP and antagonist RC-3095 (RC) blocked this effect (Fig. 4d,e and Extended Data Fig. 4b). As expected, GRP had no effect on GRPR-negative mouse (1181 and 1014) and human (501mel) cells, unless GRPR was ectopically expressed (Extended Data Fig. 4c–h).

**Fig. 4: GRPR activates YAP1 to activate the metastatic program.**

RNA-seq analysis of GRP-stimulated human and murine melanoma lines indicated the activation of key cancer-related pathways, notably anoikis resistance and invasion (Extended Data Fig. 4i–n). GRPR activation reduced apoptosis in unattached mouse and human melanoma cells, indicating resistance to anoikis. Inhibition of GRPR with RC restored sensitivity to anoikis (Fig. 4f,g and Extended Data Fig. 4o–t). GRP did not promote anoikis resistance in male ∆Ecad (1062) or female Ecad (1014) cells (Extended Data Fig. 4s,t).

GRP stimulation also promoted invasion exclusively in female GRPR-positive 1057 cells, but not in Ecad-positive (1181 male and 1039 female) or Ecad-negative male (1456) cells (Extended Data Fig. 4u–y). GRP-induced invasion in GRPR-positive lines (including 1057, 1064, MDA-MB-435S, Dauv-1 and GRPR-transduced 1014, 1181 and 501mel) was blocked by RC-3095 (Fig. 4h,i and Extended Data Fig. 4l–n,z–Af). As expected, this inhibition was absent in GRPR-negative cells treated with GRP (Extended Data Fig. 4aa–ac).

GRPR induces YAP1 through Gα_q/11

G-protein-coupled receptors signal through various Gα and downstream pathways¹². We used PamGene kinase assays to identify the main pathway activated by GRP/GRPR in melanoma cells. GRPR activation primarily triggered protein kinase C (PKC) and, to a lesser extent, PKAα and PRKX (Extended Data Fig. 5a). PKCs are typically activated by IP3/DAG signalling downstream of Gα_q/11, whereas PKA is linked to the cAMP pathway through Gα_s (ref. ¹²). RNA-seq data supported Gα_q/11 and PKC activation after GRPR stimulation (Extended Data Fig. 5b). To assess receptor coupling with Gα_q/11, we measured IP3–IP1 production after GRP treatment. GRP triggered IP1 production in GRPR-positive, but not GRPR-negative cells, confirming Gα_q/11 activation. RC-3095 blocked this response (Extended Data Fig. 5c–h).

Gα_q has been linked to YAP1 activation in non-cutaneous melanoma models¹⁴. All GRPR-induced cellular changes could be attributed to YAP1 activation^15,16,17. In human cutaneous melanoma, GRPR levels correlated with YAP1 activation scores, and mouse ∆Ecad female tumours showed a YAP1 signature that was absent in ∆Ecad male or Ecad tumours (Extended Data Fig. 5i,j). We did not observe increased YAP1 levels in ∆Ecad females or after GRPR expression in Ecad lines (Extended Data Fig. 5k,l), but YAP1 nuclear localization increased in GRP-treated cells, and less so with RC-3095 (Fig. 4j).

Transcriptomic analysis of GRPR-expressing cells (1057, 1064, 1181-Grpr, 1014-Grpr, Dauv-1 and 501mel-GRPR) showed increased YAP1 signatures and scores after GRPR activation, whereas GRPR-negative controls (1181-Ct and 1014-Ct) showed no change (Fig. 4k and Extended Data Fig. 5m–ac). YAP1 activation was also blocked by RC in GRPR-expressing cells (Extended Data Fig. 5r–w,aa–c). Finally, we confirmed that GRP/GRPR-driven invasion depends on Yap1 and Taz (Fig. 4l). In summary, GRPR activation by GRP stimulates Gα_q signalling, activating a YAP1-regulated metastasis program.

CDH1–CTNNB1–ESR1 drives female GRPR

H3K27ac ChIP–seq in mouse melanoma cell lines revealed female ∆Ecad-specific signatures linked to β-catenin (LEF1), oestrogen receptor-α (ESR1) and YAP1/TEAD (Fig. 5a). Cdh1 loss is often linked to enhanced β-catenin signalling, reflected here by increased promoter activation of β-catenin targets such as Apcdd1, Axin2, Nkd1, Notum and Sp5—primarily in ∆Ecad female melanomas—with corresponding gene expression (Extended Data Fig. 6a–j). Inhibiting Cdh1 and ectopic β-catenin expression, or reducing Apc in Ecad-expressing cells, elevated Nkd1/Axin2, Esr1 and Grpr expression (Fig. 5b,c and Extended Data Fig. 6k–n). Conversely, β-catenin inhibition through siRNA or iCRT3 decreased the expression of Nkd1, Esr1 and Grpr (Extended Data Fig. 6o), indicating that β-catenin acts upstream of Esr1 and Grpr.

**Fig. 5: *Grpr* expression in female mice by a *Cdh1*/*Ctnnb1*/*Esr1*/*Grpr* amplification loop.**

We identified a β-catenin binding site in intron 1 of Esr1, marked by H3K27ac, in female ∆Ecad cells (Extended Data Fig. 6p), indicating direct regulation. Esr1 regulation by β-catenin was validated in human melanoma using iCRT3 (Extended Data Fig. 6q,r). Inhibiting β-catenin (iCRT3 and siCtnnb1) reduced GRP-induced invasion in mouse melanoma (Extended Data Fig. 6s). Gene-set enrichment analysis (GSEA) comparison of Ecad and ∆Ecad female tumours showed that ∆Ecad tumours overexpressed genes tied to ERα activity (Extended Data Fig. 7a,b). Other sex hormone receptors (Esr2, Gper1, Ar and Pgr) were not expressed in murine melanoma, and only ∆Ecad female lines expressed both Grpr and Esr1 (Fig. 5d and Extended Data Fig. 7c). ERα protein was also detected in these lines (Fig. 5e).

ERα bound active chromatin associated with the Grpr promoter (Extended Data Fig. 7d,e), supporting direct regulation. Silencing Esr1 in Esr1⁺/Grpr⁺ melanoma reduced Grpr mRNA, whereas Esr1 overexpression increased it (Fig. 5f,g and Extended Data Fig. 7f,g), confirming that Esr1 regulates Grpr. In Ecad-expressing melanoma, CDH1 silencing in mouse and human cells raised ESR1 and GRPR mRNA (Fig. 5h and Extended Data Fig. 7h), supporting the inverse correlation between CDH1 and GRPR seen in TCGA data (Fig. 2j). ESR1 inhibition decreased GRPR mRNA, and dual CDH1/ESR1 inhibition blocked GRPR upregulation, indicating that GRPR induction on ECAD loss is at least partly ESR1 dependent (Fig. 5h and Extended Data Fig. 7h). At the protein level, CDH1 knockdown or knockout upregulated ERα (Fig. 5i and Extended Data Figs. 6t,u and 7i–k), and ESR1 silencing increased ECAD, showing reciprocal regulation (Fig. 5i and Extended Data Fig. 7l).

The ERα agonist 17β-oestradiol (E2) increased Grpr expression, whereas degrader ICI 182,780 (ICI) reduced it (Extended Data Fig. 7m). This decrease was linked to lower Grpr activity, shown by reduced IP1 production after ICI treatment (Extended Data Fig. 5c–f). Finally, the CDH1–ESR1–GRPR axis operates in a positive-feedback loop: GRPR activation represses Cdh1, and E2 reinforces this loop by activating ERα and further inhibiting Cdh1 (Fig. 5i,j and Extended Data Fig. 7m–o). Thus, ECAD loss initiates a transcriptional program amplified by feedback, explaining the high GRPR expression (Extended Data Fig. 8).

Fulvestrant inhibits invasion and metastasis

To assess the clinical relevance of oestrogen signalling in Grpr-positive melanomas, we examined the impact of oestrogen inhibition using ICI 182,780 (Fulvestrant) both in vitro and in vivo. In vitro, ICI treatment significantly reduced GRP-induced invasion of 1057 ΔEcad melanoma cells, with no effect on Grpr-negative lines derived from transgenic mice (Extended Data Fig. 9). For in vivo analysis, we used two approaches. First, 1057 ΔEcad cells were either pretreated with ICI for three days or left untreated before being injected into the tail vein of female C57BL/6J mice, simulating adjuvant therapy after primary tumour surgery. Second, untreated 1057 cells were injected, followed by Fulvestrant treatment (50 mg per kg) three hours after injection and once a week for three weeks. In both cases, mice were euthanized after 25 days and lung metastases were analysed (Fig. 5k). In the first approach, 9 of 13 mice injected with ICI-pretreated cells showed no visible metastases, compared with only 1 of 11 in the control group. In the second, 5 of 6 Fulvestrant-treated mice showed no metastases. Metastatic burden, assessed by qPCR for the 1057-specific Cre transgene, was reduced in both treatment groups (Fig. 5k). These results demonstrate that oestrogen signalling promotes the invasiveness and metastatic potential of Grpr-positive melanoma cells, and that ICI/Fulvestrant effectively suppresses these processes in vitro and in vivo.

The CDH1–GRPR axis is active in breast cancer

As shown in Fig. 2j, GRPR expression and CDH1 expression in breast tumours are inversely correlated. Given the oestrogen-dependent growth of many breast tumours¹⁸, we proposed that E-cadherin represses the ESR1–GRPR axis across tissues, including breast cancer. We observed elevated GRPR and ESR1 mRNA levels in breast tumours expressing mutated, non-functional ECAD (Extended Data Fig. 10a,b). Moreover, CDH1-mutant tumours displayed higher ERα activation scores and greater ERα positivity by immunohistochemistry (Extended Data Fig. 10c,d). Several findings supported GRPR’s dependence on ERα: the increased GRPR levels in ERα-positive compared with ERα-negative breast tumours (Extended Data Fig. 10e); the strong correlation between GRPR mRNA and ERα activation score (Extended Data Fig. 10f); and ERα binding to active regions associated with the GRPR promoter (Extended Data Fig. 10g,h). To validate the role of ERα in regulating GRPR in breast cancer, we did some experiments in MCF7 cells. CDH1 inhibition increased GRPR and ESR1 mRNA, whereas ESR1 inhibition reduced GRPR mRNA (Extended Data Fig. 10i). Notably, combined CDH1 and ESR1 silencing failed to induce GRPR expression. Activation of ERα by oestradiol increased GRPR, whereas ICI 182,780 treatment significantly decreased its expression (Extended Data Fig. 10j).

Together, these findings highlight a conserved role for this regulatory loop, indicating its broader relevance across multiple cancer types and physiological contexts.

Discussion

Our study uncovers a female-specific metastatic pathway involving E-cadherin. Reduced E-cadherin increases β-catenin transcriptional activity, elevating ESR1 and inducing GRPR transcription, which activates YAP1 and initiates a metastatic cascade. This newly identified CDH1–CTNNB1–ESR1–GRPR–YAP1 axis defines a female-specific tumour metastasis route, providing some potential therapeutic targets (Extended Data Fig. 8). E-cadherin is known to be tumour suppressor in breast and stomach cancers^19,20, in which GRP is abundantly produced and activates GRPR (Extended Data Fig. 2k,l). Thus, GRPR activation probably occurs in primary tumours in these tissues. By contrast, in other carcinomas or melanoma, GRPR is probably activated at metastatic sites by locally produced GRP (Extended Data Fig. 8). Our results indicate that E-cadherin suppresses tumour initiation in GRP-rich primary tissues and acts as to suppress metastasis in contexts in which GRP is restricted to metastatic niches, such as the lung in cutaneous melanoma²¹.

The E-cadherin–CTNNB1–ESR1–GRPR loop may be triggered by: E-cadherin alterations through mutation, methylation or epithelial–mesenchymal transition (EMT)^19,22,23; WNT/BCAT activation repressing ECAD transcription and triggering pathways including ESR1–GRPR²⁴; ERα activation through mutation, menstrual cycles, pregnancy or xenobiotics^25,26; or GRPR activation by GRP, other low-affinity ligands or receptor transactivation²⁷. These findings highlight the complex interplay of pathways in cancer progression and emphasize the need to consider them for better cancer management.

We show the sex-dependent role of E-cadherin in repressing ESR1 in human and mouse melanoma and breast cancer cells. Cadherin–catenin interactions, including those involving β-catenin, plakoglobin and p120-catenin, regulate gene expression. Without E-cadherin, β-catenin becomes transcriptionally active and upregulates ESR1. Moreover, E-cadherin loss also alters gene expression through β-catenin-independent mechanisms^28,29. Other mechanisms might be at play, given the ability of E-cadherin to independently influence gene expression³⁰. E-cadherin affects receptor tyrosine kinase signalling and reshapes chromatin through EMT, influencing CTCF expression and chromatin structure^31,32,33,34. Our ChIP–seq data support this, showing that SMARCD1-targeted genes are activated following E-cadherin loss³⁵. Given that SMARCD1 interacts with ERα, it probably helps to remodel chromatin at ERα targets such as GRPR, especially in young women³⁶. However, understanding chromatin-level effects of E-cadherin and their age and sex dependence needs further study.

A few cancers—mainly of the sex organs—are defined as hormone dependent, but others lack defined hormonal status owing to a lack of research³⁷. Our study innovatively links hormone sensitivity to tumour phenotype. In such tumours, sex hormones do not drive proliferation but influence phenotype, metastasis and therapy resistance. This hormone sensitivity probably applies to carcinomas and potentially other cancers, supported by epidemiological data (Supplementary Table 1).

We reveal that loss of E-cadherin activates a metastatic axis involving ERα, GRPR and YAP1, which are therefore potential therapeutic targets. ER modulators and aromatase inhibitors are already used as adjuvants in ER-positive breast cancer³⁸. Because ERα regulates GRPR, anti-oestrogens might suppress both and improve outcomes. However, aromatase inhibitors can cause joint and muscle pain, possibly affecting survival if treatment is stopped.

The involvement of GRPR in itch and nociception offers potential cancer and pain therapy^39,40. Despite their relevance, G-protein-coupled receptors remain underused in oncology¹². We highlight the role of GRPR in cancer progression. The GRPR antagonist RC-3095 reduced melanoma lung metastases and shows promise in breast, lung, prostate and pancreatic cancers⁴¹. Optimizing pharmacokinetics and reducing toxicity remain key challenges in the development of potent and selective antagonists⁴². It is important to develop selective GRPR antagonists with minimal side effects. Limited adult GRPR expression and mild knockout phenotypes suggest that this approach is safe^43,44. Structural data from X-ray and cryo-electron microscopy may aid the discovery of better antagonists^45,46. YAP1 is another target, although it is harder to inhibit and is potentially less safe than GRPR antagonists^47,48,49.

In summary, our findings underscore the importance of recognizing sex diversity in disease and therapy. This study introduces a sex-specific strategy that could improve cancer treatment in women in whom loss of E-cadherin leads to ERα and GRPR expression.

Methods

Cancer epidemiology

All the cancer incidence data available from the Global Cancer Observatory’s 2020 release were accessed on 24–26 October 2022. Data were collected by sex and age category. The women/men ratios were calculated from the age-standardized rates and logged for representation purposes. A fourth-order polynomial model was fitted for each cancer using prism to represent the women/men ratio dynamic. Each cancer was classified according to a calculated premenopausal variation index (PV) calculated in this way:

$${\rm{P}}{\rm{V}}={\mu }_{\text{ratio premenopause}}-\frac{({\mu }_{\text{ratio prepuberty}}+{\mu }_{\text{ratio post menopause}})}{2},$$

with µ being the average of the time period. Three categories were defined according to PV: if PV < −0.2, cancers were defined as male-biased during premenopause; if PV > 0.2, cancers were defined as female-biased during premenopause; otherwise, cancers were defined as unbiased during premenopause. The concordance between PV and the fitted curve was manually checked for each cancer.

Lists of genes associated with cancers

Lists of genes associated with melanoma (73, hsa05218), gastric cancer (150, hsa05226) and thyroid cancer (37, hsa05216) were retrieved from KEGG. The list of genes associated with breast cancer (172) was from WikiPathway (WP1984). ESR1-associated genes (317) were extracted from ESR1-associated genes from StringDB using the following parameters: search, CDH1 in human; setting, high confidence; and maximum number of interactors, 317. Lists were intersected using a Venn diagram.

Sex-hormone concentration

Plasma testosterone and oestradiol levels in males and females, categorized by age, were extracted from ref. ⁵⁰.

Cancer genomic and transcriptomic data mining

The transcriptome, copy-number alteration, mutations and corresponding clinical data of the TCGA–SKCM (n = 473), TCGA–BRCA (n = 1,215), TCGA–STAD (n = 450), TCGA–LUSC (n = 553) and TCGA–KIRC (n = 606) datasets were retrieved from the National Cancer Institute Genomic Data Commons repository using the TCGAbiolinks R package in August 2022 (ref. ⁵¹). The mRNA levels were calculated from RNA-sequencing read counts using RNA-Seq V2 RSEM and normalized to TPM. Expression of CDH1, CCND1, BRAF and KRAS was modelled over time using LOWESS smoothing with 1,860 points. Expression levels from acral melanoma were retrieved from GEO under accession number GSE162682. Survival analyses were carried out by separating the cohort into two groups according to gene expression. The threshold was set to 1 TPM, commonly considered to be the limit for sufficient protein expression of the transcript. The negative group was set to an expression 0.1 TPM or less to have clear separation in terms of expression (factor 10) from the positive group. The YAP and melanoma phenotypic state scores were obtained by averaging the expression of detailed marker genes, or for the mice, from their murine orthologues as described in ref. ⁵². The pigmentation state was defined by the expression of MITF, MLANA, TRPM1, DCT and TYR; the starved-like melanoma cell phenotype by the expression of CD36, DLX5, IP6K3, PAX3 and TRIM67; the invasive state by the expression of AXL, CYR61, TCF4, LOXL2, TNC and WNT5A; and the neural crest cell-like phenotypic state by the expression of AQP1, GFRA2, L1CAM, NGFR, SLC22A17 and TMEM176B⁵². The scoring of YAP1 activation was determined from the expression of CYR61, CTGF, TEAD4, LATS2 and CRIM1. Yap scoring was obtained by averaging the fold change of each Yap1 target. The anoikis-resistance score was calculated on the basis of the expression of S100A7A, MTPN, ATP10B, S100A8, RSAD2, RENBP, CDHR1 and CD36. The ER-activation score was calculated according to the expressions of GATA4, SDK2, EGR3, IL19, GSG1L, RSPO1, PGR, IL24 and PDZK1. Expression data from human and mouse normal tissue were downloaded from the Human Protein Atlas (https://www.proteinatlas.org/) and from the EBI expression atlas (https://www.ebi.ac.uk/gxa/home), respectively, both accessed on 7 October 2022. The anatogram was generated using the gganatogram R package⁵³.

Mice

Animal care, use and all experimental procedures were conducted in accordance with recommendations of the European Community (86/609/EEC) and European Union (2010/63/UE) and the French National Committee (87/848). Mice were housed in a specific-pathogen-free (SPF)-certified animal facility with a 12 h:12 h light:dark cycle in a temperature- and humidity-controlled room (22 ± 1 °C and 60%, respectively) with free access to water and food. Animal care and use were approved by the ethics committee of the Curie Institute in compliance with institutional guidelines. Experimental procedures were carried out under the approval of the ethics committee of the Institut Curie CEEA-IC #118 (CEEA-IC 2016-001) in compliance with international guidelines. The transgenic Tyr::CreA (B6.Cg-Tg(Tyr-cre)1Lru/J), named Tyr::Cre, Tg(Tyr-NRAS*Q61K)1Bee, Cdkn2a^tm1Rdp, B6.129-Cdh1tm2Kem/J, named Cdh1F/F, mice, also including Tyr::CreB, bcat* (Tg(Tyr-Ctnnb1/EGFP)#Lru), Pten (Pten^tm1Hwu) have been described and characterized previously in the Larue laboratory and elsewhere^{11,54,55,56,57,58}. The mouse lines were backcrossed onto a C57BL/6J background for more than ten generations. Genotyping was done according to ref. ⁵⁹ using specific primers and conditions (Supplementary Tables 9 and 10). Mice were crossed to obtain the desired genotypes. Mice were born with the expected ratio of Mendelian inheritance and no changes in gender ratios were observed. Mice were checked weekly for the appearance of new tumours. Tumour volume (V) was calculated using the formula V = (L × W²)/2, where L is the longest dimension of the tumour and W is the width perpendicular to L. Tumours were allowed to grow until reaching a volume of approximately 1 cm³. To comply with ethical guidelines, the total tumour burden per mouse was limited to 1 cm³. Mice were euthanized on reaching any predefined ethical end point. At the end of the experiments, all mice underwent autopsy to assess the presence of metastases in distant organs.

Injections, in vivo imaging and metastasis detection

Mouse and human melanoma cells (5 × 10⁵) were suspended in 200 µl PBS and injected into the tail veins of eight-week-old C57BL/6J and NSG mice, respectively. Mice were monitored daily and weighed twice weekly. Euthanasia and autopsy were performed following 20% weight loss or reaching ethical end points.

For imaging, C57BL/6J mice were shaved three days before injection and biweekly. Mice received 300 µg Xenolight d-luciferin (Revvity) intraperitoneally, followed by isoflurane anaesthesia. After 10 min, luminescence was recorded using an IVIS Spectrum system (Perkin-Elmer) with adjustable shutters focused on the thoracic region. Whole-body and thoracic luminescence were acquired for 2 min on day 0, day 1 and twice weekly. Mice were randomized after injection based on body weight and day 0 lung luminescence. Treatments (10 µg RC/DMSO or PBS/DMSO) were administered twice daily for one month, with luminescence monitoring. After euthanasia, lungs were processed as follows: right lung, incubated in 300 µg ml⁻¹ Xenolight d-luciferin (2 min) and imaged (1 min); left lung, fixed in 4% PFA (4 °C, 24 h), cryoprotected in 30% sucrose and 30% sucrose/50% OCT (48 h each), embedded in OCT, sectioned (7 µm) and stained for Dct and Grpr mRNA (RNAscope, BioTechne; probes, Mm-Dct-C1 460461, Mm-Grpr-C3 317871-C3). Imaging was done using a Leica SP8 confocal microscope.

Metastasis quantification

Right lungs were imaged using a Leica MZFLIII binocular microscope with a Scion camera. Metastases were quantified in ImageJ, with macro-metastases (bigger than 0.1 mm) and micro-metastases (0.1 mm or smaller).

ICI/Fulvestrant experiments

In vitro pre-treatment: 1057 melanoma cells (5 × 10⁵) were pretreated for 3 days with 1 µM ICI before tail-vein injection. In vivo treatment: 1057 melanoma cells (5 × 10⁵) were injected into C57BL/6J mice, followed by 50 mg Fulvestrant (Zentiva) 3 h after injection, then weekly for 3 weeks. Mice were euthanized on day 25 after injection, and lung metastases were assessed by PCR-based Cre transgene detection. DNA was extracted from lung tissue and metastasis burden was analysed statistically.

RNA extraction and transcriptomic analysis

RNA was extracted from cells and mouse tumours using the miRNeasy kit (Qiagen, 217004), according to the manufacturer’s protocol. RNA integrity (RIN) was measured using an Agilent Bioanalyser 2100 (Agilent Technologies) and an RNA nano 6000 kit (5067-1511, Agilent Technologies). Only RNA with an RNA integrity number (RIN) of more than 7 was used for analysis. This threshold led to the sequencing of 72 mouse melanoma cell lines, 36 human melanoma cell lines and 32 mouse tumours. RNA concentrations were measured using a NanoDrop (NanoDrop Technologies). RNA-sequencing libraries were prepared from 1 μg total RNA using an Illumina TruSeq Stranded mRNA library preparation kit, which allows strand-specific sequencing. PolyA selection using magnetic beads was done to focus the sequencing on polyadenylated transcripts. After fragmentation, cDNA synthesis was done and the resulting fragments used for dA-tailing, followed by ligation with TruSeq indexed adapters. The fragments were amplified by PCR to generate the final barcoded cDNA libraries (12 amplification cycles). The libraries were equimolarly pooled and subjected to qPCR quantification using the KAPA library quantification kit (Roche). Sequencing was carried out on a NovaSeq 6000 instrument (Illumina) based on a 2 × 100 cycle mode (paired-end reads, 100 bases) using an S1 flow cell to obtain approximately 35 million clusters (70 million raw paired-end reads) per sample. Reads were mapped to the mm10 mouse reference genome (gencode m13 version-GRCm38.p5) or hg38 human reference genome (gencode 42 version-GRCh38.p13) using STAR⁶⁰. STAR was also used to create the expression matrices. When applicable, expression was batch-corrected with Combat using the sva package from R⁶¹.

Differential gene-expression analysis was done using R following the DEseq2 pipeline with the DEseq2 package⁶². DEseq2 and edgeR (ref. ⁶³) (to retain only the expressed genes) algorithms were used. The packages are both available from Bioconductor (http://www.bioconductor.org) (accessed in October 2022). The threshold for significantly differentially expressed genes was set as an absolute log₂-fold change greater than 1. The volcano and correlation plots depicting the results were generated using the R package ggplot2 v.3.5.1 (ref. ⁶⁴). GSEA was done using previously published signatures, described in supplementary Table 8, and expression was obtained after DEseq2 normalization. GSEA parameters were set to 1,000 permutations per gene set. Only gene sets with a normalized enrichment score of more than 1.7 and a false discovery rate of less than 0.05 were considered.

RNA quantification by RT–qPCR

RNA (3 µg) was reverse transcribed using M‐MLV reverse transcriptase (Invitrogen), according to the manufacturer’s protocol. The newly synthesized cDNA was used as a template for qPCR with the iTaq Universal SYBR Green Supermix. Technical triplicates were used for each sample and the quantified RNA normalized against TBP (human) or Hprt (mouse) as housekeeping transcripts (Supplementary Tables 9 and 10).

Human samples

The retrospective study on lung human melanoma metastases was approved by the ethics committee. The non-opposition or consent (before or after 2004, respectively) of patients for the use of their biological material and data was obtained according to the bioethics law of 2004. We retrieved 43 tissue samples of non-treated lung melanoma metastases registered from 1999 to 2014 from the pathology files of the Bordeaux and Rennes hospital. We selected all available formalin-fixed paraffin-embedded surgical specimens of lung melanoma metastases for further immunostaining analysis.

Immunohistochemistry

Paraffin was melted at 56 °C overnight. Deparaffinization, using a Bond Dewax Solution (CAR9222, Leica), and rehydration were done with a Leica BONDTM-MAX device. Heat-induced epitope antigen retrieval was performed at 100 °C for 20 min in Bond Epitope Retrieval Solution 1 (AR9961, Leica) for GRPR or Bond Epitope Retrieval Solution 2 (1/100, AR9640, Leica) for E-cadherin. Slides were incubated in anti-GRPR antibody (SP4337P, Acris Antibodies) in Bond Primary Antibody Diluent (1/100, AR9352, Leica) overnight at 4 °C and anti-E-cadherin (1/100, NCL-L-E-Cad, Novocastra) antibody in the same diluent for 30 min at room temperature. Bond Polymer Refine Red Detection (DS9390, Leica) was used, according to the manufacturer’s specifications. Slides were counterstained with haematoxylin and cover-slipped. Images were acquired using an Axio Imager Z2 microscope. Each immunostaining was evaluated in a double-blind manner by two pathologists.

Cell lines

Mouse melanoma cell lines were established from melanomas arising in transgenic mice in the laboratory as previously described⁶⁵. The cell lines mutational landscape was determined by whole-exome sequencing. MDA-MB-435S, 624mel (often referred to as 501mel in the literature; 624mel cells are male), 888-Mel and Dauv-1 human melanoma cell lines were previously established in other laboratories^66,67,68. The human breast cancer cell line MCF-7 was previously established⁶⁹. The pGK-Luc2 vector was a gift from Catherine Tomasetto (IGBMC). In brief, the coding sequence of the luciferase reporter gene luc2 (from Photinus pyralis) was amplified by PCR from the pGL4.50[luc2/CMV/Hygro] vector (Promega, E1310) and flanking XhoI restriction sites were added. The digested PCR fragment was subcloned into the SalI site of the pLENTI PGK Blast DEST vector (plasmid 19065, Addgen). 1057-luciferase melanoma cell lines were generated by infecting parental 1057 cells with pGK-Luc2 (LL#1231). Cells were selected using 4.5 µg ml⁻¹ blasticidin for one week. Cell lines 1014 and 1181 Grpr and the corresponding controls were obtained after transfection of the murine Grpr/tGFP plasmid (1045, MG224721, Origene) and tGFP plasmid (1064, pCMV6-AC-GFP, Origene), respectively. The pSpCas9(BB)−2A-GFP (PX458) was a gift from Medhi Khaled (Institut Gustave Roussy). The annealed oligos corresponding to the gRNA sequences were ligated within the BsmbI (R0580, New England Biolabs) digested plasmid using Quick Ligase (M2200S, New England Biolabs). The oligos are listed in Supplementary Table 7. Mouse melanoma cells were transfected with either the plasmid targeting the 5′ side of the second exon or the plasmids targeting the 5′ and the 3′ side of exon 2 using lipofectamine 2000 (11668019, Invitrogen). Two days later, GFP-positive cells were sorted and one cell per well was seeded in 96-well plates. Cells were cultured until first passage when half of the cells were collected for DNA extraction. The status of Grpr was assessed by PCR. Transfection of β-catenin was performed using CMV::bcat* (777) or mock control (empty pcDNA3, 297). Cells were transfected with 2–4 µg of plasmid or 100 pmol of siRNA and lipofectamine 2000, following the manufacturer’s protocol. Transfected 1014 and 1181 cells were selected using 25 or 150 µg ml⁻¹ geneticin, respectively. Then, 501mel-GRPR and 501mel-Ct were generated by infecting cells with the pLV-Hygro-CMV-Grpr-EGFP (LL1272, 1VB191126-1286xxe, Vector Builder) or pLV-Hygro-CMV-EGFP (LL1271b, VB191126-1289cfv, Vector Builder) plasmids, respectively. For siRNA knockdown, cells were transfected with 100 pmol siRNA (Supplementary Table 11) using Lipofectamine 2000 (Invitrogen) following the manufacturer’s instructions. Inducible piSMART shRNA plasmids (V3SM11253-231787949 for the shGrpr and VSC11655 for the non-targeting shRNA, Dharmacon) were infected in 1057 melanoma cells, and positive cells were selected using 1.2 µg ml⁻¹ puromycin. Before experiments, cells were treated for three days with doxycyclin or mock. For CRISPR, the control was generated using 1.2 µg scramble gRNA + 6 µg Cas9-RFP and 1.2 µg 3′ gRNA + 6 µg Cas9-GFP, and CDH1 was generated using 1.2 µg 5′ gRNA + 6 µg Cas9-RFP and 1.2 µg 3’ gRNA + 6 µg Cas9-GFP. The gRNA and Cas9 constructs (ALT-R optimization) were obtained from Integrated DNA Technologies. Both RNP complexes were transfected into 888-Mel melanoma cells using lipofectamine CRISPRMAX (cmax00008, Invitrogen). GFP⁺/RFP⁺ cells were sorted the next day and cultured in phenol red-free media with 100 nM E2 for four days. The gRNA sequences are shown in Supplementary Table 9.

Murine and human melanoma cell lines were grown in Ham’s F12 medium and RPMI 1640, respectively, supplemented with 10% FCS (10270106, Gibco) and 1% PS. The breast cancer cell line MCF7 was grown in DMEM-F12 supplemented with 10% FCS and 1% PS. All cell lines were maintained at 37 °C in a humidified atmosphere with 5% CO₂. Cells cultured without phenol red were supplemented with 2 nM glutamine. The genetic status and level of expression of key genes of these cell lines are presented in Supplementary Table 7.

Whole-exome sequencing

We used 2 million mouse melanoma cells. The DNA was extracted using the DNeasy Blood and Tissues kit (69504, Qiagen). Library preparation was done using the SureSelect XT Mouse All Exon Kit (Agilent) followed by high-throughput sequencing on an Illumina NovaSeq 6000 instrument (Illumina). Analyses were performed with the European galaxy instance. We aligned the fastq to the mouse mm10 genome using BWA. Duplicates were removed using the MarkDuplicates (v.3.1.1.0) function from Picard. Tracks were visualized using IGV.

Cell growth and clonogenic assay

Six-well tissue-culture plates were seeded with 3 × 10⁵ melanoma cells in complete medium. After 24 h, the medium was replaced by low-serum medium (0.5% for murine melanoma cell lines and 1% for human melanoma cells) and the cells were incubated for a further 18 h before stimulation with 10 nM GRP (4011670 bachem) and/or 1 µM RC (R9653, Sigma-Aldrich) for 48 h. The plates were trypsinized just after stimulation or 48 h later and the cells were counted. For the MTT assay, 10,000 cells were seeded per well in 96-well plates. After 24 h, cells were starved overnight and then treated with 10 nM GRP and/or 1 µM RC for 48 h in low-serum medium (0.5% FBS). Next, MTT (M5655, Sigma-Aldrich) was added to the wells to a final concentration of 0.5 mg ml⁻¹ and the plates were incubated for 3 h. The medium was removed and formazan crystals dissolved in 200 µl DMSO. The absorbance was read at 570 nM using a LUMIstar Omega luminometer (BMG Labtech). All growth experiments were done using three technical replicates and three biological replicates. For the clonogenic assay, six-well tissue-culture plates were seeded with 500 cells in complete medium. After 10 days (20 days for 1181), colonies were fixed with 4% PFA for 15 min and stained with 10% crystal violet in ethanol for 20 min and counted in images using ImageJ software. Experiments were done in triplicate.

Anoikis assay

Six-well plates were coated with poly-HEMA to prevent cell attachment to the well surface. Cells were seeded in low-serum medium containing 10 nM GRP and/or 1 µM RC. Then, 48 hours after cell seeding, cells were collected and washed with ice-cold PBS twice and resuspended in annexin V binding buffer (556454, BD Biosciences) and incubated at room temperature in the dark with 7-amino-actinomycin D (7-AAD, 559925, BD Biosciences) and annexin-V for 15 min. Annexin V was coupled to FITC (556420, BD Biosciences) for the 1057 and MDA-MB-435S cell lines and to PE (556421, BD Biosciences) for the other cell lines. Cells were sorted using a FACS LSRFortessa (BD Biosciences) to determine the percentage of annexin V- and/or 7-AAD-positive cells using a 488 nm laser for annexin V-FITC and annexin V-PE and a 675 nm laser for7-AAD. All quantification was performed using Flow-Jo.

Matrigel invasion assay

Matrigel invasion assays were done in transwell plates with 8.0-µm pores (353097, Falcon) coated with 100 µl of 200 µg ml⁻¹ Matrigel. Depending on the experiments, cells were seeded directly or 24 h after transfection in low-serum medium (0.5% FCS for murine cells and 1% for human cell lines) with 10 nM GRP, 1 µM RC-3095, and/or 1 µM IC 182,780, or in 10% FCS. Then, 24 hours after stimulation, inserts were washed with PBS and non-invading cells were removed. Cells in the inserts were fixed in methanol at −20 °C overnight. The inserts were rinsed and the membrane carefully removed using a sharp scalpel. The membrane was mounted in prolong glass DAPI (1.5 µg ml⁻¹). Assays were performed in triplicate and automatic counting of invading cells was done using an Image/Fiji macro (https://doi.org/10.5281/zenodo.14509394). Sequences of siRNAs are given in Supplementary Table 11.

Western blot analysis

Whole-cell lysates were prepared and analysed as described previously⁶⁵. The primary antibodies used were: E-cadherin (610182, BD Transduction Laboratories, dilution 1/1,000); ERα (MA1-27107, Invitrogen, dilution 1/500); YAP (14074, Cell Signaling, dilution 1/1,000); β-actin (A5441, Sigma, dilution 1/5,000); and vinculin (4650, Cell Signaling, dilution 1/5,000). All raw data are provided in Supplementary Fig. 1.

Kinase assay

Serine/threonine and tyrosine kinase activity were determined using STK PamChips (87102 PamGene International). All assays were done according to the manufacturer’s protocol⁷⁰. In brief, cells were seeded to 70% confluence and starved overnight the day after. Then, the medium was replaced by medium containing vehicle (0.1% DMSO) or 10 nM GRP and/or 1 µM RC and the cells incubated for 15 min. Cells were rinsed twice in PBS and then lysed in M-Per buffer (78503, ThermoFisher Scientific) containing Halt Phosphatase Inhibitor (78428, ThermoFisher Scientific) and Halt Protease Inhibitor (78437, ThermoFisher Scientific), both diluted 1:100. The lysates were immediately snap-frozen. Kinase activity was determined using Pamstation PS12 and 1 µg protein for the STK chips, according to the manufacturer’s protocol. The data were analysed using BioNavigator software (PamGene International), batch corrected using ComBat⁷¹ and normalized using VSN⁷². Kinase activity was assessed using the 2018 version of the UKA tool using basic parameters (Scan rank from 4 to 12, 500 permutations, 90% homology, equivalent weight for each database, minimal prediction score of 300). The UKA tool infers the active kinase from the differentially phosphorylated peptides using databases and predicted interactions (PhosphoNet database).

IP1 detection

IP1 levels were quantified using the HTRF IP-one Gq detection kit (62IPAPEB, Revvity) following the manufacturer’s protocol. We treated 4,000 cells for 30 min at 37 °C in StimB buffer with 10 nM GRP, 1 µM RC-3095, 1 µM ICI-182,780 or DMSO. FRET signals (620 nm and 665 nm, excitation at 385 nm) were measured using a Clariostar plate reader (BMG Labtech).

Immunofluorescence

Cells were seeded on coverslips and cultivated until 100% confluence was achieved. After 24 h of starvation in 0.5% FCS for murine melanoma cell lines and 1% FCS for human melanoma, the medium was complemented with 10 nM GRP and/or 1 µM RC for 1 h. Cells were fixed with 4% PFA for 15 min and blocked with 5% normal goat serum and 0.3% Triton X-100 in PBS. Cells were incubated overnight at 4 °C with anti-YAP D8H1X antibody (14074, cell Signaling, 1/100 dilution) in 1% BSA and 0.3% Triton X-100 followed by incubation for 1.5 h at room temperature with goat anti-rabbit Alexa fluor 594 (A-11012, Invitrogen, 1/500 dilution). Coverslips were mounted with Prolong Gold containing 1.5 µg ml⁻¹ DAPI (P36934, Invitrogen). Images were acquired using an inverted SP8 Leica confocal microscope (Leica Microsystem). YAP1 localization was evaluated using ImageJ Software.

Plasma stability assay

Each compound was diluted in mouse plasma to a final concentration of 1 μM and incubated at 37 °C for 2 h. The reaction was stopped by the addition of 2.5 volumes of ice-cold acetonitrile. Liquid chromatography–mass spectrometry was performed on the supernatant in multiple reaction-monitoring mode (LC–MS/MS). The percentage of the remaining test compound relative to that present at the start was determined by monitoring the peak area.

Metabolic stability assay

Each compound was diluted to a final concentration of 1 µM in 100 mM phosphate buffer (pH 7.4) containing 0.5 mg ml⁻¹ mouse liver microsomes, 1 mM NADPH regenerating system and 1 mM MgCl₂ and incubated at 37 °C for 1 h. At various time points (0, 2, 10, 20, 40 and 60 min), one volume of ice-cold acetonitrile was added and the supernatants were analysed by LC–MS/MS. To obtain a stability curve, the percentage of remaining test compound at each time point was determined by monitoring the peak area. The half-life was estimated from the slope of the initial linear range of the logarithmic curve of the remaining compound (percentage) against time, assuming first-order kinetics.

LC–MS/MS

Analyses were done on a Shimadzu 8030 LC–MS instrument. Chromatographic separations were done at 40 °C using a 2.6-µm C18 Kinetex column (50 × 2.1 mm; Phenomenex). The mobile phase flow rate was set to 0.5 ml min⁻¹ and the following program applied for the elution: 0 min, 5% B; 0–1.2 min, 5–95% B; 1.2–1.4 min, 95% B; 1.4–1.42 min, 95-5% B; and 1.42–2.8 min, 5% B (solvent A, 0.05% formic acid in water; solvent B, acetonitrile). The injection volume was 1 µl. The mass spectrometer was interfaced with the liquid-chromatography system using an electrospray ion source. The nitrogen nebulizing gas flow was set to 1.5 l min⁻¹ and the drying gas flow to 15 ml min⁻¹. The interface voltage was set to 4,500 V. The temperature of the block heater was maintained at 400 °C and the desolvation line at 250 °C. Argon was used as the collision gas at 230 kPa. The transitions in positive mode were m/z 585.1 → 204.0, 221.0 for PD176252, and m/z 369.5 → 144.1, 110.1 for RC.

ChIP–seq

Mouse melanoma cell lines were grown in normal medium until reaching 60 × 10⁶ cells at 70% confluency. Cells were fixed in 10 ml of 0.4% PFA for 10 min at room temperature. Crosslinking reactions were stopped by adding 1 ml of 2 M glycine (pH 8.0). Cells were resuspended in 1 ml PBS and sonicated until DNA was fragmented to an average of 400 bp. For H3K27ac ChIP–seq in mouse cell lines, 100 µg chromatin was then incubated for 4 hours at 4 °C with 40 µl protein G sepharose beads (17-0618-02, VWR Chemicals) blocked with tRNA from yeast (AM7119, Invitrogen) and BSA in ChIP dilution buffer. Supernatant was pipetted to a new tube and incubated overnight on a rotating wheel at 4 °C with 10 µg anti-H3K27ac antibody. The chromatin was incubated for 1 hour at 4 °C on a rotating wheel with 20 µl blocked protein G sepharose beads. The supernatant was removed and beads were washed with the ChIP low-salt buffer, twice with the ChIP high-salt buffer, twice with LiCl buffer and twice with TE buffer. Chromatin was eluted by incubating the beads twice with 250 µl elution buffer for 15 min on a rocking plate at room temperature. Then, 20 µl of 5 M NaCl was added to the chromatin and the mix was incubated at 65 °C overnight to reverse crosslinks. Proteins were degraded by incubating the chromatin with 10 µl TrisHCl (pH 6.8), 10 µl of 0.5 M EDTA and 20 µg proteinase K for 1 hour at 42 °C. DNA was extracted adding 500 µl phenol:chloroform (1:1, pH 7–8) followed by extraction of the supernatant with 500 µl chloroform. The supernatant was removed and DNA was precipitated by adding 50 µl of 3 M sodium acetate, 15 µg Glycoblue and 1 ml ethanol, and incubated at 4 °C for 4 hours. Pellets were washed with ethanol (70%), dried and finally resuspended in 25 µl water. ChIP-buffer compositions are detailed in Supplementary Table 12. Libraries were prepared from input and immunoprecipitated DNA using the Illumina TruSeq ChIP library preparation kit, according to the manufacturer’s protocol. In brief, 2–3 ng of DNA was subjected to subsequent steps of end-repair, dA-tailing and ligation of TruSeq RNA UD index Illumina adapters. After a final PCR amplification step (15 cycles), the 13 resulting bar-coded libraries were equimolarly pooled and quantified using a qPCR method (KAPA library quantification kit, Roche). Sequencing was done on a NovaSeq 6000 instrument from Illumina using paired-end 2 × 100 bp, to obtain around 60 million clusters (120 million raw paired-end reads) per sample. Raw files were uploaded on the European instance of galaxy (usegalaxy.eu) for further processing⁷³. Reads were aligned to the GRCm38 mm10 reference genome using bowtie2 (ref. ⁷⁴). The same analysis process was used for the following raw ChIP–seq files from the literature, downloaded from GEO: ERα in mouse uterus (GSM894054; ref. ⁷⁵); β-catenin in nephron progenitor cells (GSM980186; ref. ⁷⁶); H3K27ac in MCF7 (GSM2175784; ref. ⁷⁷); ATAC-seq (GSM2645717; ref. ⁷⁸); ERα in MCF7 (GSM798434; ref. ⁷⁹); and β-catenin in hESC (GSM1579346; ref. ⁸⁰). For human ChIP–seq, reads were aligned against the human GRCh38 hg38 genome. The 3D chromatin interactions were downloaded from GSE207828 (ref. ⁸¹) for mouse and from ENCODE⁸² (www.encodeproject.org, accession number ENCFF804SET) and ref. ⁸³ for human. All ChIP–seq alignments and genome interactions were visualized using IGV. ChIP–seq data were annotated using chipseeker⁸⁴.

Pharmacological targeting of ERα

For hormone depletion, FCS was stripped using the dextran-coated charcoal method. In brief, Norit activated charcoal (C6241-5G, Sigma-Aldrich, final concentration of 0.25%) and Dextran T-70 (31390, Sigma-Aldrich, final concentration of 0.0025%) in 0.25 M sucrose/1.5 mM MgCl₂/10 mM HEPES (H4034, Sigma-Aldrich), pH 7.4, were incubated overnight at 4 °C. The volume equivalent to the volume of the FCS to strip was pipetted into a new 50 ml tube and centrifugated at 500g for 10 min to remove the supernatant. The FCS was incubated with the activated charcoal for 12 h at 4 °C. The treated FCS was then filtered through a 0.22-µM pore filter to ensure sterility and mixed with the appropriate phenol red-free medium and PS. Cells were collected using phenol red-free trypsin and starved for oestrogen using phenol red-free/10% stripped FCS for four days. Cells were then stimulated with β-oestradiol (2824, Tocris Bioscience) or ICI 182,780 (1047, Tocris Bioscience) for 72 h in phenol red-free/5% stripped FBS.

Statistical analysis

Cell culture-based experiments were performed in at least biological triplicates and validated three times as technical triplicates. The significance of the effects was calculated using the Mann–Whitney test or Student’s t-test for the comparison of two groups for non-parametric or parametric situations, respectively. When more than two groups were compared, Kruskal–Wallis or ANOVA was used, according to the parametricity of the data. After each Kruskal–Wallis or ANOVA test, results were adjusted for multiple comparisons using the Dunn and Tukey adjustments, respectively. Categorical data were compared using Fisher’s exact test when two groups were compared or, otherwise, a Chi-squared test. The significance of the difference between Kaplan–Meier curves was calculated using a log-rank test. Data are represented as mean ± s.d. unless otherwise indicated in the figure legend. All P-values are reported as computed by Prism 10. P < 0.05 was considered significant; NS, not significant; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All sequencing data generated with this manuscript were deposited on the relevant platform. RNA-seq and ChIP–seq data were deposited on GEO at the National Center for Biotechnology Information under the SuperSeries GSE218588. That includes: mouse tumour RNA-seq with accession number GSE218532; mouse cell line RNA-seq with accession number GSE218586; and human cell line data with accession number GSE218530. The ChIP–seq data are available with accession number GSE237500. Whole-exome sequencing of the mouse melanoma cell lines are available from SRA at bioproject PRJNA904253 (https://dataview.ncbi.nlm.nih.gov/object/40767225). Kinase-assay raw data were deposited on Mendeley at https://doi.org/10.17632/nwkpyr2nmh.1. The following datasets from the literature were used: ChIP–seq of ERα in mouse (GSM894054) and human (GSM798434), of β-catenin in mouse (GSM980186) and human (GSM1579346) and H3K27Ac in human MCF7 (GSM2175784). ATAC-seq data from the MCF7 cell line originate from GSM2645717. The 3D chromatin interactions were downloaded from GSE207828 for mouse and from ENCODE ENCFF804SET (https://www.encodeproject.org/experiments/ENCSR549MGQ/) and GSE52457 for human. Expression data from acral melanoma are from GSE190113. TCGA datasets were accessed through TCGABiolinks or CBioPortal (https://www.cbioportal.org/). All data have been aligned to the human reference genome GRCh38 (hg38 gencode 42 version GRCh38.p13) and the mouse reference genome mm10 gencode 13 version GRCm38.p5. All other data are available from the corresponding author upon reasonable request. Source data are provided with this paper.

Change history

14 July 2025
A Correction to this paper has been published: https://doi.org/10.1038/s41586-025-09353-9

References

Dong, M. et al. Sex differences in cancer incidence and survival: a pan-cancer analysis. Cancer Epidemiol. Biomarkers Prev. 29, 1389–1397 (2020).
Article PubMed Google Scholar
Ye, Y. et al. Sex-associated molecular differences for cancer immunotherapy. Nat. Commun. 11, 1779 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Vellano, C. P. et al. Androgen receptor blockade promotes response to BRAF/MEK-targeted therapy. Nature 606, 797–803 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Zhao, Y., Wang, X., Liu, Y., Wang, H.-Y. & Xiang, J. The effects of estrogen on targeted cancer therapy drugs. Pharmacol. Res. 177, 106131 (2022).
Article CAS PubMed Google Scholar
Rubin, J. B. et al. Sex differences in cancer mechanisms. Biol. Sex Differ. 11, 17 (2020).
Article CAS PubMed PubMed Central Google Scholar
Sung, H. et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. 71, 209–249 (2021).
PubMed Google Scholar
Cook, M. B. et al. Sex disparities in cancer incidence by period and age. Cancer Epidemiol. Biomarkers Prev. 18, 1174–1182 (2009).
Metcalfe, A. et al. Incidence of pregnancy-associated cancer in two Canadian provinces: a population-based study. Int. J. Environ. Res. Public Health 18, 3100 (2021).
Article PubMed PubMed Central Google Scholar
Lee, Y. Y. et al. Incidence and outcomes of pregnancy-associated cancer in Australia, 1994–2008: a population-based linkage study. BJOG 119, 1572–1582 (2012).
Article CAS PubMed PubMed Central Google Scholar
Benyi, E. et al. Risks of malignant and non-malignant tumours in tall women treated with high-dose oestrogen during adolescence. Horm. Res. Paediatr. 82, 89–96 (2014).
Article CAS PubMed Google Scholar
Ackermann, J. et al. Metastasizing melanoma formation caused by expression of activated N-RasQ61K on an INK4a-deficient background. Cancer Res. 65, 4005–4011 (2005).
Article CAS PubMed Google Scholar
Raymond, J. H., Aktary, Z., Larue, L. & Delmas, V. Targeting GPCRs and their signaling as a therapeutic option in melanoma. Cancers 14, 706 (2022).
Article CAS PubMed PubMed Central Google Scholar
Farshidfar, F. et al. Integrative molecular and clinical profiling of acral melanoma links focal amplification of 22q11.21 to metastasis. Nat. Commun. 13, 898 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Yu, F.-X. et al. Mutant Gq/11 promote uveal melanoma tumorigenesis by activating YAP. Cancer Cell 25, 822–830 (2014).
Article CAS PubMed PubMed Central Google Scholar
Verfaillie, A. et al. Decoding the regulatory landscape of melanoma reveals TEADS as regulators of the invasive cell state. Nat. Commun. 6, 6683 (2015).
Article ADS CAS PubMed Google Scholar
Zhang, X. et al. The Hippo pathway oncoprotein YAP promotes melanoma cell invasion and spontaneous metastasis. Oncogene 39, 5267–5281 (2020).
Article CAS PubMed Google Scholar
Zhao, B. et al. YAP activation in melanoma contributes to anoikis resistance and metastasis. Exp. Biol. Med. (Maywood) 246, 888–896 (2021).
Article CAS PubMed Google Scholar
Sommer, S. & Fuqua, S. A. Estrogen receptor and breast cancer. Semin. Cancer Biol. 11, 339–352 (2001).
Article CAS PubMed Google Scholar
Berx, G. & Van Roy, F. The E-cadherin/catenin complex: an important gatekeeper in breast cancer tumorigenesis and malignant progression. Breast Cancer Res. 3, 289–293 (2001).
Article CAS PubMed PubMed Central Google Scholar
Guilford, P. et al. E-cadherin germline mutations in familial gastric cancer. Nature 392, 402–405 (1998).
Article ADS CAS PubMed Google Scholar
Uhlén, M. et al. Proteomics. Tissue-based map of the human proteome. Science 347, 1260419 (2015).
Article PubMed Google Scholar
Cano, A. et al. The transcription factor Snail controls epithelial–mesenchymal transitions by repressing E-cadherin expression. Nat. Cell Biol. 2, 76–83 (2000).
Article CAS PubMed Google Scholar
Venza, M. et al. DNA methylation-induced E-cadherin silencing is correlated with the clinicopathological features of melanoma. Oncol. Rep. 35, 2451–2460 (2016).
Article CAS PubMed Google Scholar
Larue, L. & Delmas, V. The WNT/Beta-catenin pathway in melanoma. Front. Biosci. 11, 733–742 (2006).
Article CAS PubMed Google Scholar
Avberšek, M., Žegura, B., Filipič, M., Uranjek-Ževart, N. & Heath, E. Determination of estrogenic potential in waste water without sample extraction. J. Hazard. Mater. 260, 527–533 (2013).
Article PubMed Google Scholar
Robinson, D. R. et al. Activating ESR1 mutations in hormone-resistant metastatic breast cancer. Nat. Genet. 45, 1446–1451 (2013).
Article CAS PubMed PubMed Central Google Scholar
Liu, X.-Y. et al. Unidirectional cross-activation of GRPR by MOR1D uncouples itch and analgesia induced by opioids. Cell 147, 447–458 (2011).
Article CAS PubMed PubMed Central Google Scholar
Herzig, M., Savarese, F., Novatchkova, M., Semb, H. & Christofori, G. Tumor progression induced by the loss of E-cadherin independent of β-catenin/Tcf-mediated Wnt signaling. Oncogene 26, 2290–2298 (2007).
Article CAS PubMed Google Scholar
Jeanes, A., Gottardi, C. J. & Yap, A. S. Cadherins and cancer: how does cadherin dysfunction promote tumor progression? Oncogene 27, 6920–6929 (2008).
Article CAS PubMed PubMed Central Google Scholar
Larue, L. et al. A role for cadherins in tissue formation. Development 122, 3185–3194 (1996).
Article CAS PubMed Google Scholar
Morali, O. G. et al. IGF-II induces rapid β-catenin relocation to the nucleus during epithelium to mesenchyme transition. Oncogene 20, 4942–4950 (2001).
Article CAS PubMed Google Scholar
Qian, X., Karpova, T., Sheppard, A. M., McNally, J. & Lowy, D. R. E-cadherin-mediated adhesion inhibits ligand-dependent activation of diverse receptor tyrosine kinases. EMBO J. 23, 1739–1748 (2004).
Article CAS PubMed PubMed Central Google Scholar
Elangovan, A. et al. Loss of E-cadherin induces IGF1R activation and reveals a targetable pathway in invasive lobular breast carcinoma. Mol. Cancer Res. 20, 1405–1419 (2022).
Article CAS PubMed PubMed Central Google Scholar
Johnson, K. S. et al. CTCF expression and dynamic motif accessibility modulates epithelial-mesenchymal gene expression. Cancers 14, 209 (2022).
Article CAS PubMed PubMed Central Google Scholar
Centore, R. C., Sandoval, G. J., Soares, L. M. M., Kadoch, C. & Chan, H. M. Mammalian SWI/SNF chromatin remodeling complexes: emerging mechanisms and therapeutic strategies. Trends Genet. 36, 936–950 (2020).
Article CAS PubMed Google Scholar
Hsiao, P.-W., Fryer, C. J., Trotter, K. W., Wang, W. & Archer, T. K. BAF60a mediates critical interactions between nuclear receptors and the BRG1 chromatin-remodeling complex for transactivation. Mol. Cell. Biol. 23, 6210–6220 (2003).
Article CAS PubMed PubMed Central Google Scholar
Emons, G. Hormone-dependent cancers: molecular mechanisms and therapeutical implications. Cells 12, 110 (2022).
Article PubMed PubMed Central Google Scholar
Rižner, T. L. & Romano, A. Targeting the formation of estrogens for treatment of hormone dependent diseases–current status. Front. Pharmacol. 14, 1155558 (2023).
Article PubMed PubMed Central Google Scholar
Saeki, A., Yamanaka, H., Kobayashi, K., Okubo, M. & Noguchi, K. Analgesic effect of gastrin-releasing peptide in the dorsal horn. Mol. Pain 18, 17448069221108965 (2022).
Article CAS PubMed PubMed Central Google Scholar
Polgár, E. et al. Grpr expression defines a population of superficial dorsal horn vertical cells that have a role in both itch and pain. Pain 164, 149–170 (2023).
Article PubMed Google Scholar
Patel, O., Shulkes, A. & Baldwin, G. S. Gastrin-releasing peptide and cancer. Biochim. Biophys. Acta 1766, 23–41 (2006).
CAS PubMed Google Scholar
Schwartsmann, G. et al. A phase I trial of the bombesin/gastrin-releasing peptide (BN/GRP) antagonist RC3095 in patients with advanced solid malignancies. Invest. New Drugs 24, 403–412 (2006).
Article CAS PubMed Google Scholar
Hampton, L. L. et al. Loss of bombesin-induced feeding suppression in gastrin-releasing peptide receptor-deficient mice. Proc. Natl Acad. Sci. USA 95, 3188–3192 (1998).
Article ADS CAS PubMed PubMed Central Google Scholar
Chen, Z.-F. A neuropeptide code for itch. Nat. Rev. Neurosci. 22, 758–776 (2021).
Article CAS PubMed PubMed Central Google Scholar
Peng, S. et al. Structures of human gastrin-releasing peptide receptors bound to antagonist and agonist for cancer and itch therapy. Proc. Natl Acad. Sci. USA 120, e2216230120 (2023).
Article CAS PubMed PubMed Central Google Scholar
Li, C. et al. Molecular recognition of itch-associated neuropeptides by bombesin receptors. Cell Res. 33, 184–187 (2023).
Article CAS PubMed Google Scholar
Pobbati, A. V. & Hong, W. A combat with the YAP/TAZ-TEAD oncoproteins for cancer therapy. Theranostics 10, 3622–3635 (2020).
Article CAS PubMed PubMed Central Google Scholar
He, Z., Li, R. & Jiang, H. Mutations and copy number abnormalities of Hippo pathway components in human cancers. Front. Cell Dev. Biol. 9, 661718 (2021).
Article PubMed PubMed Central Google Scholar
Mills, K. R., Misra, J. & Torabifard, H. Allosteric modulation of the YAP/TAZ-TEAD interaction by palmitoylation and small-molecule inhibitors. J. Phys. Chem. B 128, 3795–3806 (2024).
Article CAS PubMed Google Scholar
Deltourbe, L. G. et al. Steroid hormone levels vary with sex, aging, lifestyle, and genetic. Sci. Adv. 11, eadu6094 (2025).
Colaprico, A. et al. TCGAbiolinks: an R/Bioconductor package for integrative analysis of TCGA data. Nucleic Acids Res. 44, e71 (2016).
Article PubMed Google Scholar
Rambow, F. et al. Toward minimal residual disease-directed therapy in melanoma. Cell 174, 843–855 (2018).
Article CAS PubMed Google Scholar
Maag, J. L. V. gganatogram: An R package for modular visualisation of anatograms and tissues based on ggplot2. F1000Research https://doi.org/10.12688/f1000research.16409.2 (2018).
Delmas, V. et al. β-Catenin induces immortalization of melanocytes by suppressing p16^INK4a expression and cooperates with N-Ras in melanoma development. Genes Dev. 21, 2923–2935 (2007).
Article CAS PubMed PubMed Central Google Scholar
Boussadia, O., Kutsch, S., Hierholzer, A., Delmas, V. & Kemler, R. E-cadherin is a survival factor for the lactating mouse mammary gland. Mech. Dev. 115, 53–62 (2002).
Article CAS PubMed Google Scholar
Delmas, V., Martinozzi, S., Bourgeois, Y., Holzenberger, M. & Larue, L. Cre-mediated recombination in the skin melanocyte lineage. Genesis 36, 73–80 (2003).
Article CAS PubMed Google Scholar
Serrano, M. et al. Role of the INK4a locus in tumor suppression and cell mortality. Cell 85, 27–37 (1996).
Article CAS PubMed Google Scholar
Lesche, R. et al. Cre/loxP-mediated inactivation of the murine Pten tumor suppressor gene. Genesis 32, 148–149 (2002).
Article CAS PubMed Google Scholar
Hamm, M. et al. BRN2 is a non-canonical melanoma tumor-suppressor. Nat. Commun. 12, 3707 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Article CAS PubMed Google Scholar
Leek, J. T., Johnson, W. E., Parker, H. S., Jaffe, A. E. & Storey, J. D. The sva package for removing batch effects and other unwanted variation in high-throughput experiments. Bioinformatics 28, 882–883 (2012).
Article CAS PubMed PubMed Central Google Scholar
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
Article PubMed PubMed Central Google Scholar
Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010).
Article CAS PubMed Google Scholar
Wickham, H. ggplot2: Elegant Graphics for Data Analysis (Springer, 2016).
Petit, V. et al. C57BL/6 congenic mouse NRAS^Q61K melanoma cell lines are highly sensitive to the combination of Mek and Akt inhibitors in vitro and in vivo. Pigment Cell Melanoma Res. 32, 829–841 (2019).
Article CAS PubMed Google Scholar
Cailleau, R., Olivé, M. & Cruciger, Q. V. Long-term human breast carcinoma cell lines of metastatic origin: preliminary characterization. In Vitro 14, 911–915 (1978).
Article CAS PubMed Google Scholar
Marincola, F. M. et al. Loss of HLA haplotype and B locus down-regulation in melanoma cell lines. J. Immunol. 153, 1225–1237 (1994).
Article CAS PubMed Google Scholar
Rambow, F. et al. New functional signatures for understanding melanoma biology from tumor cell lineage-specific analysis. Cell Rep. 13, 840–853 (2015).
Article CAS PubMed PubMed Central Google Scholar
Soule, H. D., Vazguez, J., Long, A., Albert, S. & Brennan, M. A human cell line from a pleural effusion derived from a breast carcinoma. J. Natl Cancer Inst. 51, 1409–1416 (1973).
Article CAS PubMed Google Scholar
Chirumamilla, C. S. et al. Profiling activity of cellular kinases in migrating T-cells. Methods Mol. Biol. 1930, 99–113 (2019).
Article CAS PubMed Google Scholar
Johnson, W. E., Li, C. & Rabinovic, A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics 8, 118–127 (2007).
Article PubMed Google Scholar
Motakis, E. S., Nason, G. P., Fryzlewicz, P. & Rutter, G. A. Variance stabilization and normalization for one-color microarray data using a data-driven multiscale approach. Bioinformatics 22, 2547–2553 (2006).
Article CAS PubMed Google Scholar
The Galaxy Community. The Galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2022 update. Nucleic Acids Res. 50, W345–W351 (2022).
Article Google Scholar
Langmead, B., Wilks, C., Antonescu, V. & Charles, R. Scaling read aligners to hundreds of threads on general-purpose processors. Bioinformatics 35, 421–432 (2019).
Article CAS PubMed Google Scholar
Hewitt, S. C. et al. Research resource: whole-genome estrogen receptor α binding in mouse uterine tissue revealed by ChIP-seq. Mol. Endocrinol. 26, 887–898 (2012).
Article CAS PubMed PubMed Central Google Scholar
Park, J.-S. et al. Six2 and Wnt regulate self-renewal and commitment of nephron progenitors through shared gene regulatory networks. Dev. Cell 23, 637–651 (2012).
Article CAS PubMed PubMed Central Google Scholar
Rhie, S. K. et al. Identification of activated enhancers and linked transcription factors in breast, prostate, and kidney tumors by tracing enhancer networks using epigenetic traits. Epigenetics Chromatin 9, 50 (2016).
Article PubMed PubMed Central Google Scholar
Nair, S. J. et al. Phase separation of ligand-activated enhancers licenses cooperative chromosomal enhancer assembly. Nat. Struct. Mol. Biol. 26, 193–203 (2019).
Article CAS PubMed PubMed Central Google Scholar
Ross-Innes, C. S. et al. Differential oestrogen receptor binding is associated with clinical outcome in breast cancer. Nature 481, 389–393 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Estarás, C., Benner, C. & Jones, K. A. SMADs and YAP compete to control elongation of β-catenin:LEF-1-recruited RNAPII during hESC differentiation. Mol. Cell 58, 780–793 (2015).
Article PubMed PubMed Central Google Scholar
Perovanovic, J. et al. Oct1 cooperates with the Smad family of transcription factors to promote mesodermal lineage specification. Sci. Signal. 16, eadd5750 (2023).
Article CAS PubMed PubMed Central Google Scholar
Luo, Y. et al. New developments on the Encyclopedia of DNA Elements (ENCODE) data portal. Nucleic Acids Res. 48, D882–D889 (2020).
Article CAS PubMed Google Scholar
Dixon, J. R. et al. Chromatin architecture reorganization during stem cell differentiation. Nature 518, 331–336 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Yu, G., Wang, L.-G. & He, Q.-Y. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics 31, 2382–2383 (2015).
Article CAS PubMed Google Scholar

Download references

Acknowledgements

We thank D. Metzger for providing ESR1 expression vector and comments; A. Bellacosa, D. Duteil, M. Dutertre, C. Goding, S. Marullo and E. Steingrismsson for conversations; F. Beermann, R. Kemler and M. Seano for providing the Tg(Tyr-NRAS*Q61K)1Bee, B6.129-Cdh1tm2Kem/J and Cdkn2a^tm1Rdp mouse lines, respectively; J. Groten, M. Bel, E. Van Breemen, S. Rangarajan and J. Lebens for PamGene Technology; the staff at the Institut Curie responsible for the animal colony (especially P. Dubreuil and C. Lantoine), histology (S. Leboucher), FACS (C. Lasgi) and PICT-IBiSA imaging (C. Messaoudi and L. Besse); and D. Delmas, J.-C. Delmas, M. Larue and R. Larue for their constant support. This work was supported by FRM EQU202103012599 and the Institut National Du Cancer (INCa; 2023-1-PL BIO-03-ICR-1), Institut Curie, INSERM and CNRS. J.H.R. had a fellowship from PSL and LNCC. M.W. had a fellowship from DIM SEnT. F.L. had a fellowship from MENRT, SFD and LNCC. High-throughput sequencing was done by the ICGex NGS platform of the Institut Curie, supported by grants ANR-10-EQPX-03 (Equipex) and ANR-10-INBS-09-08 (France Génomique Consortium) from the Agence Nationale de la Recherche (Investissements d’Avenir programme), by the ITMO-Cancer Aviesan (Plan Cancer III) and by the SiRIC-Curie programme (SiRIC grant INCa-DGOS-465 and INCa-DGOS Inserm_12554). Data management, quality control and primary analysis were done by the Bioinformatics platform of the Institut Curie.

Author information

These authors contributed equally: Lionel Larue, Véronique Delmas

Authors and Affiliations

Normal and Pathological Development of Melanocytes, Institut Curie, PSL Research University, INSERM U1021, Orsay, France
Jérémy H. Raymond, Zackie Aktary, Marie Pouteaux, Valérie Petit, Flavie Luciani, Maria Wehbe, Lionel Larue & Véronique Delmas
Université Paris-Saclay, CNRS UMR 3347, Orsay, France
Jérémy H. Raymond, Zackie Aktary, Marie Pouteaux, Valérie Petit, Flavie Luciani, Maria Wehbe, Lionel Larue & Véronique Delmas
CNRS UMS3286, Plateforme de Chimie Biologique Intégrative de Strasbourg, Strasbourg University/Strasbourg Drug Discovery and Development Institute (IMS), Strasbourg, France
Patrick Gizzi & Claire Bourban
Laboratory of Preclinical Investigation, Department of Translational Research, Institut Curie, PSL University Paris, Paris, France
Didier Decaudin & Fariba Nemati
IGBMC, CNRS UMR7104, INSERM U1258, Université de Strasbourg, Illkirch, France
Igor Martianov, Irwin Davidson & Catherine-Laure Tomasetto
Ludwig Institute for Cancer Research, University of Oxford, Oxford, UK
Richard M. White
CNRS UMR9187, INSERM U1196, Chemistry and Modeling for the Biology of Cancer, Institut Curie, Université PSL, Orsay, France
Florence Mahuteau-Betzer
Université Paris-Saclay, CNRS UMR 9187, INSERM U1196, Orsay, France
Florence Mahuteau-Betzer
Service de pathologie CHU de Bordeaux et Eq. Translational Research on Oncodermatology and Orphean Skin Diseases (TRIO2) BoRdeaux Institute of onCology (BRIC) UMR 1312, INSERM/Université de Bordeaux, Bordeaux, France
Béatrice Vergier

Authors

Jérémy H. Raymond
View author publications
Search author on:PubMed Google Scholar
Zackie Aktary
View author publications
Search author on:PubMed Google Scholar
Marie Pouteaux
View author publications
Search author on:PubMed Google Scholar
Valérie Petit
View author publications
Search author on:PubMed Google Scholar
Flavie Luciani
View author publications
Search author on:PubMed Google Scholar
Maria Wehbe
View author publications
Search author on:PubMed Google Scholar
Patrick Gizzi
View author publications
Search author on:PubMed Google Scholar
Claire Bourban
View author publications
Search author on:PubMed Google Scholar
Didier Decaudin
View author publications
Search author on:PubMed Google Scholar
Fariba Nemati
View author publications
Search author on:PubMed Google Scholar
Igor Martianov
View author publications
Search author on:PubMed Google Scholar
Irwin Davidson
View author publications
Search author on:PubMed Google Scholar
Catherine-Laure Tomasetto
View author publications
Search author on:PubMed Google Scholar
Richard M. White
View author publications
Search author on:PubMed Google Scholar
Florence Mahuteau-Betzer
View author publications
Search author on:PubMed Google Scholar
Béatrice Vergier
View author publications
Search author on:PubMed Google Scholar
Lionel Larue
View author publications
Search author on:PubMed Google Scholar
Véronique Delmas
View author publications
Search author on:PubMed Google Scholar

Contributions

L.L. and V.D. conceived the work, designed the experimental set-up, did data analysis and supervised the study. J.H.R., Z.A., V.P., I.M., I.D. and V.D. did biochemical and cell-biological experiments and data analysis. J.H.R., M.P., F.L., M.W., P.G., C.B., D.D., F.N., B.V. and V.D. did in vivo experiments and data analysis. J.H.R. and V.P. did in silico analysis. C.-L.T., R.M.W. and F.M.-B. provided critical resources. J.H.R. and L.L. produced the graphical representation of the data. L.L. and V.D. acquired funding to support the project. The manuscript was written by J.H.R., L.L. and V.D. The manuscript was reviewed by Z.A., V.P., I.D. and R.M.W. All authors approved and contributed to the final manuscript.

Corresponding author

Correspondence to Lionel Larue.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature thanks Svenja Meierjohann, Alpha Yap and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data figures and tables

Extended Data Fig. 1 Between puberty and the menopause, the incidence of cancer is higher in women than in men.

a,b Crude incidence rates (cases/100,000) for all cancers (a) and cutaneous malignant melanoma (b) in the world population in 2020. Data are presented for each sex and as the average of both sexes. c, Cancers with a positive female/male curve peak and PV > 0.20 (PV = premenopausal variation) are highlighted in red, cancers with a negative curve peak and PV < −0.20 are highlighted in blue, and cancers without a curve peak are colored in white (see Supplementary Table 1). d-f, Examples of the three categories: (d) gastric cancer (positive peak), (e) liver cancer (negative peak), and (f) brain/CNS cancer (no peak). Dashed lines represent 4th order polynomial fits. g, Median testosterone concentration by age in women and men. h, CDH1 expression in melanoma by NRAS or BRAF drivers (TCGA-SKCM), Chi-square test. *p ≤ 0.05. i,j, Micro- (i) and macro- (j) metastasis quantification by Ecad and sex. Each dot is an indenpendant mouse. Proportion were assessed by a Chi-square and metastasis counts by two-sided t-test adjusted. Data are representated as mean ± SD.

Source data

Extended Data Fig. 2 Grpr transcription activation in ∆Ecad female melanoma and endogenous agonist expression in the lung.

a, RT-qPCR measured Grpr mRNA in male and female Ecad and ∆Ecad mouse melanoma. b, H3K27ac ChIP-seq track for Grpr in mouse melanoma cell lines by sex and Ecad status, aligned to GRCm38 mm10 reference genome. c, RNAscope of mouse lungs showing Grpr (green) and Dct (red - as a melanoma marker) in Ecad and ΔEcad male and female transgenic mice; ΔEcad females with lung metastasis show abundant Grpr expression. d, Grpr mRNA levels in Tyr::NRAS^Q61K female melanomas background with different genetic backgrounds. Tyr::CreA is located on the X chromosome (denoted by ‘A’), while Tyr::CreB is located on an autosome (denoted by ‘B’). ‘-’ represents the absence of the Cre gene in the genome. Ecad +/+ is indicated by ‘+’, EcadF/F by ‘-‘, Ink4a +/+ by ‘+’, and Ink4a +/− by ‘−’. Similarly, Pten +/+ is represented by ‘+’, and PtenF/+ by ‘−’. e, Invasive score in TCGA-SKCM melanomas with GRPR expression (Spearman coefficient r = 0.35). f-h, Differential gene expression (fold change) correlated with phenotypic scores: (f) NCSC-like, (g) pigmented, and (h) SMC. Grpr is highlighted in red. i,j, Overall Survival (OS) (i) and Progression-Free Survival (PFS) (j) Kaplan-Meier curves for TCGA-SKCM based on GRPR expression (low/absent ≤ 0.1 TPM and expressed > 1 TPM). k-m, mRNA GRP levels in humans (k,l) and mice (m); abundant in human and mouse lung. Human data sourced from the Human Protein Atlas, mouse data from the EBI Expression Atlas. Anatogram was created with gganatogram. Comparisons significance was assessed by two-sided Mann-Whitney adjusted for multiple comparisons with a Benjamini–Hochberg test. Score significance was assessed by ANOVA adjusted with a Tukey post-test. Survival was assessed by Log-Rank. Data are represented as mean ± SD. Box plot represent the median and the 25–75 percentiles, the whiskers represent the minimum and the maximum. ≥ 3 independent biological replicates were performed per experiment.

Source data

Extended Data Fig. 3 RC-3095 suppresses lung melanoma growth and metastases.

a, Lung metastases observed 30 days post-tail-vein-injection: 12/13 mice injected with ∆Ecad cells developed metastasis, whereas no metastasis was observed in 0/6 mice with Ecad cells, b, Time to reach 1 cm³ tumor volume in neck-graft reimplantations of male and female Ecad and ΔEcad melanoma in male and female C57BL/6 J mice, respectively. c, DNA and RNA sequencing of Grpr in 1057 mouse melanoma cells showing X inactivation. The blue vertical line corresponds to the active X chromosome, while the red line the inactive X chromosome. d, Grpr knockout strategy via exon 2 deletion in 1057 mouse melanoma cells. e, Genotyping of Grpr deletion clones before and after CRISPR editing. (1) corresponds to Grpr wt/wt, and (2) corresponds to Grpr wt/ΔEx2. DDW = double distilled water. Par = parental. f, Clone obtention efficiency in ΔEcad mouse melanoma cells with CRISPR-targeted gRNAs (Chi-square test). g, Grpr expression in clonal populations (1) before and (2) after CRISPR editing. The crossed-out number “3” indicates that we were unable to obtain double homozygous knockout mutants Grpr/Cdh1. Statistical analysis was performed using t-test. h, Relative mRNA Grpr expression in doxycycline-inducible shRNA-expressing 1057 melanoma cells. i, Relative mRNA Ccn1 expression following GRP induction in 1057 melanoma cells. j,k, GRPR mRNA levels in mouse (j) and human (k) melanoma cell lines, normalized in TPM and log-transformed for visualization. Ct = control – cells transfected with an empty expression vector. l,m, Fold change in CCN1 mRNA levels after 4 h 10 nM GRP. Fold changes were calculated for each biological replicate based on TPM-normalized CCN1 expression. Statistical analysis was performed using one-way ANOVA with a Tukey’s post test. n,o, Relative growth (au) of 1057 melanoma cells in response to GRP and/or GRPR inhibitors (n) PD-176252 (PD) and (o) RC-3095 (RC) assessed via MTT assay. p,q, Ex vivo stability of RC-3095 (red) and PD-176252 (blue) in murine plasma (p) and liver microsomes (q). Statistical analysis was performed using t-test. r,s, Thorax radiance (r) and mean weight (s) 30 min post-injection in RC-treated and vehicle-treated groups after randomization. t, Representative IVIS images of C57BL/6 J mice intravenously injected with 1057-Luc melanoma cells from day (d) 0 to 31. At d31, a signal was detected in additional organs, and dissection revealed metastases in the liver, adrenal glands, and ovary outside of the lungs at this time. Scale bar = 1 cm. u, Luminescence imaging of lungs after euthanasia and dissection of RC-treated (#6 to #10) or untreated (#1 to #5) mice. Scale bar= 1 cm. Lum: luminescence, Dir: direct. Proportions were evaluated by Fisher’s exact test. Comparisons significance was assessed by two-sided Mann-Whitney adjusted in case of multiple comparisons with a Benjamini–Hochberg test. Expression data were assessed by a two-sided Student T-test on the log-normalised values. Data are represented as mean ± sd. Box plot represent the median and the 25–75 percentiles, the whiskers represent the minimum and the maximum. ≥ 3 independent biological replicates were performed per experiment.

Source data

Extended Data Fig. 4 GRPR enhances melanoma cell growth and invasiveness.

a, Clonogenic growth of Grpr^neg melanoma cell line 1014: parental, control, and after exogenous Grpr expression. b-h, Effect of Grpr stimulation on growth of Grpr^pos and Grpr^neg mouse and human melanoma cell lines (b) 1064, (c) 1181-Ct, (d) 501mel-Ct, (e) 1014-Ct, (f) 1181-Grpr, (g) 501mel-GRPR, and (h) 1014-Grpr. Cell quantification presented as mean ± SD. 1181 and 501mel are GRPR^neg cells. i-k, Anoikis resistance score based on RNA-seq TPM-normalized data for GRPR^pos murine (i) 1057 and (j) 1064, and human (k) Dauv-1 melanoma cell lines. l-n, Invasive score from RNA-seq TPM-normalized data for GRPR^pos murine 1057 (l), 1064 (m), and human Dauv-1 (n) melanoma cell lines. o-t, Resistance to anoikis assays showing apoptotic (apop.) cell percentage after 48 h without matrix attachment. 1181-Ct (o), 501mel-Ct (p), 1181-Grpr (q), 501mel-GRPR (r), 1014 (s), and 1062 (t) cells treated with 10 nM GRP, 1 μM RC-3095 or no treatment. u-af, Invasion assays indicating the number (nb) of invading cells (invad.). u, Invasion assay of 1057 mouse melanoma cells, with DAPI-stained nuclei after crossing the Matrigel layer. Scale bar = 500 µm. v-y, Quantification of invading cells after GRP induction (10 nM) with or without RC-3095 in male Ecad 1181 (v), female Ecad 1039 (w), male ΔEcad 1456 (x), and female ΔEcad 1057 (y). z-af, Invasion assay in Matrigel® (200 µg/mL) for 1064 (z), 1014-Ct (aa), 1181-Ct (ab), 501mel-Ct (ac), 1014-Grpr (ad), 1181-Grpr (ae), and 501mel-GRPR (af) murine and human melanoma cells. Cells were starved overnight and treated with 10 nM GRP, 1 µM RC-3095 and/ or 10% FCS for 24 h (u-y, z-af), and for 48 h (b-h, o-t). Clonogenic assays were evalued by a Kruskall-wallis corrected by a Dunn’s post test. Effect of the GRP induction was assessed by two-sided Mann-Whitney adjusted for multiple comparisons with a Benjamini–Hochberg test. Data shown as mean ± SD. ≥ 3 independent biological replicates were performed per experiment.

Source data

Extended Data Fig. 5 GRPR activation by GRP promotes the YAP1 transcriptional program.

a, Activation of kinases following GRPR activation by 10 nM GRP in GRPR^pos melanoma cell line 1057, measured 15 min post-induction using the Pamgene® Ser/Thr kinase PamChip (STK). b, Gene set enrichment analysis (GSEA) of Gαq (left) and PKC (right) activation signatures 4 h after stimulating GRPR^pos 1057 cells with 10 nM GRP. Gene expression, determined by RNA-seq, was normalized with DEseq2 before analysis. c-h, IP1 levels, indicating Gαq/11 activation, measured in mouse melanoma cell lines after 10 nM GRP stimulation, with or without treatment with the GRPR inhibitor RC-3095 (c-h) or pre-treatment with the ERα inhibitor ICI-182,780 (c-f). i, YAP1 activation score correlated with GRPR expression in TCGA-SKCM melanoma data. TPM: transcripts per million. j, GSEA of YAP1 activation signature in mouse primary melanoma tumors: ∆Ecad vs Ecad female tumors and male tumors. RNA-seq data normalized with DEseq2 before GSEA. k, Western blot showing Yap1 protein levels in mouse melanoma cell lines. Actin as loading control. Data from one representative experiment of three biological replicates. l, Yap1 mRNA expression in melanoma cell line with or without ectopic Grpr and controls not expressing Grpr (1014-Ct, 1014-Grpr, 1181-Ct, 1181-Grpr). m-q, GSEA of YAP1 activation in murine melanoma cell lines, 4 h after 10 nM GRP stimulation. RNA-seq data normalized using DEseq2 before GSEA. r-w, Yap1 score after GRP induction in murine melanoma cells treated with vehicle, 10 nM GRP, 1 µM RC, or both. x-z, GSEA of YAP1 activation in human melanoma cell lines, 4 h after 10 nM GRP stimulation. RNA-seq data normalized using DEseq2 before GSEA. aa-ac, Yap1 score in human melanoma cells treated with vehicle, 10 nM GRP, 1 µM RC, or both. GRP effect on the IP1 level was assessed by two-sided Mann-Whitney adjusted for multiple comparisons with a Benjamini–Hochberg test. Significance of the expression was assessed by a tow-sided t-test on the log-normalized data. Significance of the scores was evaluated on the Zscore data by ANOVA adjusted by a Tukey’s test. Data are presented as mean ± SD.

Source data

Extended Data Fig. 6 Loss of Ecad in melanoma cells induces β-catenin activity and Esr1 promoter activation.

a-e, ChIP-seq tracks for H3K27ac in male and female NRAS murine melanoma cells with or without Ecad. Known β-catenin target genes: (a) Apcdd1, (b) Axin2, (c) Nkd1, (d) Notum, and (e) Sp5. f-j, Expression of β-catenin targets (Apcdd1, Axin2, Nkd1, Notum, and Sp5) in male and female mouse melanoma cells with or without not Ecad. k,l, Expression of Ctnnb1 mRNA as determined by RT-qPCR in Ecad mouse 1014 melanoma cells transfected with (k) with siScr or siCdh1 in the presence of β-catenin and (l) pcDNA3 (bcat -) or with a β-catenin expression vector (bcat +) known as bcat*. m, Effect of β-catenin activation via siCdh1 on Cdh1, Axin2, and Esr1 expression in three mouse melanoma cell lines. n, Effect of β-catenin activation using siApc on Apc, Axin2, and Esr1 expression in three mouse melanoma cell lines. o, RT-qPCR analysis of Nkd1, Axin2, Esr1, and Grpr mRNA levels in ΔEcad mouse 1057 melanoma cells treated with siRNA targeting Ctnnb1 and the pharmacological β-catenin inhibitor iCRT3 (10 μM, 48 h). p, β-catenin binding to Esr1 in nephron precursor cells with significant peaks (called by MACS2) indicated. Genomic locations are based on the mm10. q,r, RT-qPCR (q and Western blot (r) of Esr1 levels in Daju human melanoma cells treated with the β-catenin inhibitor iCRT3 (80 μM, 48 h). s, Invasion assays of Grpr-expressing mouse melanoma cells (1057), treated 24 h with siScr, siCtnnb1, iCRT3 and/or GRP. t,u RT-qPCR of CDH1 levels in human melanoma cells treated with siCDH1 (t). Western blot analysis of ERα protein levels in human melanoma cells after CDH1 knockdown-three biological replicates (u). Relative ERα levels are shown quantified in the histogram. Comparisons significance was assessed by two-sided Mann-Whitney with a Benjamini–Hochberg test. Expression data were assessed by a two-sided Student T-test on the log-normalised values. Zscores were analysed by ANOVA corrected by a Tukey’s test. Data are represented as mean ± SD. ≥ 3 independent biological replicates were performed for each experiment.

Source data

Extended Data Fig. 7 Esr1 induces Grpr expression and cell invasion.

a,b, GSEA comparing ERα expression (a) and activation (b) signatures between murine ∆Ecad and Ecad female primary melanoma. c, Quantification of Esr1 mRNA level by qRT-PCR and normalized to Hprt expression in various cell lines. d, Localization of the Grpr region on the mouse X chromosome, with three upstream regions ERα binding regions (Peak1-3, green boxes) identified through 3D interaction e, ChIP-seq tracks for H3K27ac in male and female NRAS murine melanoma cells with or without Ecad in Peak1-3 regions, showing ERα binding peaks from mouse uterus ChIP-seq data (GSM894054). Stars indicate peaks identified by MACS2. f,g, Impact of ESR1 silencing (f) and overexpression (g) on GRPR expression in human melanoma cell lines. Statistical analysis was performed using t-test. h, Consequences of CDH1 and/or ESR1 silencing on GRPR expression in CDH1^pos human melanoma cells. i, Schematic representation of the CDH1 knockout strategy. Exons 6 to 10 were deleted using two distinct guide RNAs (gRNA1 and gRNA2). j, mRNA expression levels of CDH1, ESR1, and GRPR in the 888-Mel cell population transfected with gRNA targeting CDH1. Statistical analysis was performed using t-test. k, Western blot analysis (left) and quantification (right) of ERα and E-cadherin (ECAD) protein levels in the 888-Mel cell population transfected with gRNA targeting CDH1. β-Actin and Vinculin (Vinc.) were used as loading controls. Statistical analysis was performed using a t-test. l, quantification of the western blot analysis of E-cadherin and ERα protein levels in mouse 1014 melanoma cells 48 h after knockdown with siScr (control), siCdh1 (targeting E-cadherin), or siESR1 (targeting ERα) presented in Fig. 5i. m, Impact of ERα activation (100 nM β-estradiol, E2) or degradation (1 µM ICI 182,780) on GRPR expression in ESR1^pos 1057 melanoma cells after estrogen starvation for four days. n, Quantification of the western blot analysis of (ECAD) expression in 501mel cells, expressing or not GRPR, in the presence or absence of 10 nM GRP presented in Fig. 5j. o, mRNA CDH1 levels in melanoma cells expressing GRPR, measured in the presence or absence of GRP and/or RC-3095. TMM = Trimmed Mean of M-values. For expression data, significance was evaluated by two-sided Student T-test (two groups) or ANOVA adjusted by a Tukey’s test (multiple groups) on the log-normalised data. Data are represented as mean ± SD. ≥ 3 independent biological replicates were performed per experiment. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Source data

Extended Data Fig. 8 Graphical abstract.

a, In primary melanoma, the loss of CDH1 activates β-catenin signaling, subsequently triggering the upregulation of ESR1. Consequently, ERα stimulates the expression of GRPR, particularly potent in the presence of estrogen (E2), especially during the period between puberty and menopause. Activation of GRPR by GRP subsequently reduces CDH1 expression, thereby reinforcing this loop and resulting in elevated levels of GRPR expression. ERa and GRPR inhibitors have the potential to affect this loop of regulation. b, ECAD^neg/GRPR^pos cancer cells grow in tissues expressing naturally GRP including lung, breast and gastric tissue. c, ECAD^neg/GRPR^pos melanoma cells gain the ability to disseminate through the bloodstream to distant organs, notably the lungs, where GRP is produced in abundance. Within the lungs, the interaction between GRP and GRPR initiates pro-metastatic signaling in melanoma cells.

Extended Data Fig. 9 GRP and ERα modulate melanoma cell invasion and in a sex- and E-cadherin–dependent manner.

a, Invasion assay of 1057 mouse melanoma cells, quantified by counting DAPI-stained nuclei that migrated through the Matrigel layer. Scale bar = 500 µm. b-e, Quantification of invading cells in various melanoma cell lines: male Ecad 1181 (b), female Ecad 1039 (c), male ΔEcad 1456 (d), and female ΔEcad 1057 (e) treated with 10 nM GRP, 1 μM RC-3095, and/or 1 μM ICI. f, Western blot analysis of ERα protein levels in mouse melanoma cells treated or not with ICI (1 μM, 24 h) with actin as loading control. g, Histograms showing ERα quantification from three independent biological experiments. GRP-induction significance was assessed by two-sided Mann-Whitney adjusted with a Benjamini–Hochberg test. Quantification was evaluated by ANOVA on the log transformed data. Data are represented as mean ± SD. ≥ 3 independent biological replicates were performed per experiment.

Source data

Extended Data Fig. 10 The CDH1/CTNNB1/ESR1/GRPR Axis Is Active in Breast Carcinomas.

a,b, Expression of GRPR (a) and ESR1 (b) mRNA in breast carcinoma from TCGA-BRCA data based on CDH1 genetic status. c, ERα activation score according to the CDH1 genetic status in breast cancer tumors from TCGA data. d, Quantification of ERα positivity in breast carcinomas from the TCGA-BRCA cohort by IHC based on CDH1 mutation. e, Expression of GRPR based on ERα status defined by TCGA pathologists using immunohistochemistry in human breast carcinomas from the TCGA-BRCA studies. f, Correlation between GRPR expression and the ERα activity score calculated from the same breast tumors from the TCGA-BRCA studies. g, Localization of the GRPR region on the human X chromosome (top) and identification of five regions with chromatin openness (H3K27ac) and ERα binding revealed by 3D interaction with the GRPR locus in MCF7 breast cancer cells. h, Genome browser view of ChIP-seq tracks for H3K27ac, ATAC-seq, and E2-activated ERα in MCF7 breast cancer cells on the five regions. Three regions/peaks (1-3) are conserved between humans and mice. i, Effects of CDH1 and/or ESR1 silencing on GRPR expression in CDH1-positive MCF7 human breast cancer cells. (see legend of Fig. 5h for more information). j, Effect of ESR1 stimulation (17β-estradiol, 100 nM) and inhibition (ICI 182,780, 1 µM) on mRNA GRPR expression in MCF7 cells. Expression were compared with a Student T-test (two groups) or anova with Tukey’s posttest (multiple groups) performed on the log-transformed data. Proportions were analyzed by Chi-square and correlation by Pearson. All panels, ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. The relative mRNA levels in (a, c, d, and g) are presented as Log10 (TPM + 0.01). ≥ 3 independent biological replicates were performed per experiment. Box plot represent the median and the 25–75 percentiles, the whiskers represent the minimum and the maximum.

Source data

Supplementary information

Supplementary Fig. 1

Original source images for all data obtained by electrophoretic separation. Data are organized b figure, with all replicates included. The data displayed in each Figure and Extended Figure are indicated with red boxes.

Reporting Summary

Supplementary Tables

Supplementary Tables 1–12

Peer Review File

Source data

Source Data Figs. 1, 2, 3, 4 and 5, and Source Data Extended Data Figs. 1, 2, 3, 4, 5, 6, 7, 9 and 10.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Raymond, J.H., Aktary, Z., Pouteaux, M. et al. Targeting GRPR for sex hormone-dependent cancer after loss of E-cadherin. Nature 643, 801–809 (2025). https://doi.org/10.1038/s41586-025-09111-x

Download citation

Received: 09 October 2023
Accepted: 07 May 2025
Published: 11 June 2025
Issue date: 17 July 2025
DOI: https://doi.org/10.1038/s41586-025-09111-x

This article is cited by

Sex differences in skin-cancer risk are linked to oestrogen levels
- Svenja Meierjohann
Nature (2025)

Subjects

Abstract

Similar content being viewed by others

Main

E-cadherin as an oestrogen regulator

E-cadherin loss drives sex-biased metastasis

Ecad loss upregulates GRPR expression

GRPR drives lung metastases and is targetable

GRPR fuels key metastasis pathway

GRPR induces YAP1 through Gαq/11

CDH1–CTNNB1–ESR1 drives female GRPR

Fulvestrant inhibits invasion and metastasis

The CDH1–GRPR axis is active in breast cancer

Discussion

Methods

Cancer epidemiology

Lists of genes associated with cancers

Sex-hormone concentration

Cancer genomic and transcriptomic data mining

Mice

Injections, in vivo imaging and metastasis detection

Metastasis quantification

ICI/Fulvestrant experiments

RNA extraction and transcriptomic analysis

RNA quantification by RT–qPCR

Human samples

Immunohistochemistry

Cell lines

Whole-exome sequencing

Cell growth and clonogenic assay

Anoikis assay

Matrigel invasion assay

Western blot analysis

Kinase assay

IP1 detection

Immunofluorescence

Plasma stability assay

Metabolic stability assay

LC–MS/MS

ChIP–seq

Pharmacological targeting of ERα

Statistical analysis

Reporting summary

Data availability

Change history

14 July 2025

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding author

Ethics declarations

Competing interests

Peer review

Peer review information

Additional information

Extended data figures and tables

Supplementary information

Source data

Rights and permissions

About this article

Cite this article

Share this article

This article is cited by

Search

Quick links

GRPR induces YAP1 through Gα_q/11